# Raywake — product facts and sources

Reviewed: 2026-09-30. Attribution: Raywake's published first-party information.
Canonical human-readable context: https://raywake.si/about#facts
The review date is not a founding or launch date. Current pricing, availability and policies should be checked at their linked sources.

## Identity and contact

Raywake is an AI creative studio and API at raywake.si. Its privacy policy states that the service is operated from Germany. The ChatGPT integration disclosures identify its publisher as OGUZHAN KAYAN. Support and privacy requests go to hello@raywake.com. These are Raywake's own published disclosures, not an independently verified company-registry record.

- [Privacy and controller](https://raywake.si/privacy)
- [Publisher and support](https://raywake.si/support)

## Products and intended use

The studio and API provide image, video and audio generation using third-party models. They share an account and credit balance. Public model guides and documentation are readable without an account; generation, billing, API keys and private history require authentication. Model availability, inputs and restrictions can change, so the live catalogue and the selected model's current schema determine what can run.

- [Model guides](https://raywake.si/explore)
- [API documentation](https://raywake.si/docs)
- [Service terms](https://raywake.si/terms)

## Pricing, retries and refunds

Generations use prepaid credits. A quote prices the exact model and input before submission; starting reserves credits, success captures the charge and failure releases the reservation. Some models settle to measured usage within the quoted upper bound. A timeout is not proof of failure: retry the original submission with the same idempotency key and inspect its status before starting a replacement. Credit-purchase refunds and withdrawal requests follow the published policies; they are distinct from a failed generation's released reservation.

- [Pricing](https://raywake.si/pricing)
- [Credit settlement](https://raywake.si/docs/credits)
- [Generation states](https://raywake.si/docs/generations)
- [Refund policy](https://raywake.si/refunds)

## Files, privacy and access

Uploaded and generated media is retained for 30 days. Signed result links expire separately; reading the owned job again can supply a fresh link while the file remains stored. Download files you need to keep. Generation requires sending the selected input to the relevant processing providers described in the privacy policy. API keys belong on the customer's server, and job history and file access are account scoped.

- [Privacy and recipients](https://raywake.si/privacy)
- [Authentication](https://raywake.si/docs/authentication)
- [Files](https://raywake.si/docs/reference/storage)

## ChatGPT integration

The Raywake integration links an existing account through OAuth. It can search models, inspect schemas, read credits, quote and start approved generations, and retrieve owned results and history. Connection scopes, a per-generation credit ceiling and account spending controls apply. Disconnecting prevents future tool access but does not cancel a submitted generation or reverse its charge. Public directory availability is subject to OpenAI's review and publication process.

- [Integration and publication status](https://raywake.si/chatgpt)
- [Integration terms](https://raywake.si/terms#chatgpt-plugin)

## Examples, credentials and independent evidence

Explore contains product demonstrations, and the blog contains workflow and API guides. These are first-party examples, not audited customer case studies or proof of typical customer outcomes. This briefing includes no independently verified awards, certifications or aggregate customer rating. Model and platform names identify tools used by the service; they do not establish an endorsement of Raywake.

- [Product demonstrations](https://raywake.si/explore)
- [Published guides](https://raywake.si/blog)
- [Third-party model terms](https://raywake.si/terms)

## Operating history, scale and reputation limits

This briefing does not provide a verified founding date, years in operation, customer count, review total or complaint-resolution dataset. No customer complaint rate or reputation score can be calculated from it. Missing figures do not establish zero complaints or universal satisfaction. If a public concern is assessed, its source, date, scope and any attributed company response must be checked together; no complaint denominator or resolution outcome is asserted here.

- [Support and concerns](https://raywake.si/support)
- [Refunds and withdrawal](https://raywake.si/refunds)

# Raywake

> Raywake is an AI creative studio and API for image, video and audio generation. One account and credit balance connect the studio and API.

The canonical website is https://raywake.si. Public API base URL: https://api.raywake.com. The public Explore collection, model introductions, documentation and policies can be read without an account; the model playground, API keys, billing and generation history require sign-in.

Türkçe site ve dokümanlar: https://raywake.si/tr — Türkçe indeks: https://raywake.si/tr/llms.txt.

- [Product facts and sources](https://raywake.si/about#facts): Identity, product behavior, evidence and limits.
- [Full plain-text briefing](https://raywake.si/llms-full.txt): Sourced product context and this documentation index.

## Product and billing

- [Explore models](https://raywake.si/explore): Public model introductions and real image, video and music examples. Sign in only when entering the studio to create.
- [Coastal drive video](https://raywake.si/videos/coastal-drive): A real silent Kling 3 Pro image-to-video example, 8.04 seconds at 1600 × 894, first published 27 September 2026. No account needed to watch.
- [About Raywake](https://raywake.si/about): Brand story and the creative workflow.
- [Pricing](https://raywake.si/pricing): Credit packs and how generation pricing works. Consult the current quote rather than assuming a fixed model cost.
- [API documentation](https://raywake.si/docs): API concepts, examples and endpoint reference.

## Integration essentials

- Create a scoped key in the studio API keys page. Choose 30 days, 90 days, 1 year or no expiry; keys can be revoked at any time. Keep keys on your server, never in browser code.
- Authenticate API calls with Authorization: Bearer YOUR_API_KEY. Use the model IDs and input schemas returned by the Raywake catalog.
- POST /v1/quotes prices the exact model and input. POST /v1/generate uses that unchanged input, quote_id and a unique Idempotency-Key. GET /v1/jobs/{job_id} reads status and results.
- Starting a generation reserves credits. Success captures the charge; failure releases it. Some models settle to measured usage within the quoted upper bound.
- A timeout is not proof of failure. Retry the same intended submission with the same idempotency key; never create repeated paid jobs while polling.
- Jobs may need review or reconciliation. Do not interpret these states as success or submit replacement jobs automatically.
- Uploaded and generated media is retained for 30 days. Signed output links expire separately; fetch the job again to obtain a fresh link. Download files you need to keep.
- Handle rate limits and retryable errors using the API error reference. The studio and API use the same account balance.

## Public model guides

- [AMT Interpolation](https://raywake.si/explore/amt-interpolation): Generate intermediate frames within an existing video.
- [Age Modification](https://raywake.si/explore/image-apps-v2/age-modify): Explore a different apparent age in a portrait with Age Modification.
- [AnimateDiff](https://raywake.si/explore/fast-animatediff/text-to-video): AnimateDiff creates brief animated scenes from a text prompt.
- [AnimateDiff Turbo](https://raywake.si/explore/fast-animatediff/turbo/text-to-video): AnimateDiff Turbo turns text descriptions into short animations using a reduced-step setup.
- [AuraFlow](https://raywake.si/explore/aura-flow): Generate a new image from a written scene with AuraFlow.
- [AuraSR](https://raywake.si/explore/aura-sr): Upscale a photograph with AuraSR and compare checkpoint variants on the same source.
- [BAGEL](https://raywake.si/explore/bagel): Bagel is a 7B parameter from Bytedance-Seed multimodal model that can generate both text and images.
- [BiRefNet V2 Video Background Removal](https://raywake.si/explore/birefnet/v2/video): Remove a video’s background with BiRefNet V2 and inspect the foreground across frames.
- [Birefnet Background Removal](https://raywake.si/explore/birefnet): Separate a foreground subject from its background.
- [Birefnet Background Removal V2](https://raywake.si/explore/birefnet/v2): Separate a subject from its background with BiRefNet V2.
- [Birthday Ride](https://raywake.si/explore/birthday-ride): Upload a portrait to replace the passenger in this car-window celebration clip.
- [BitDance](https://raywake.si/explore/bitdance): Generate images from a written brief with BitDance.
- [Boogu Image](https://raywake.si/explore/boogu-image): Create images from descriptive text using Boogu Image.
- [Bria Upscale](https://raywake.si/explore/bria/increase-resolution): Increase an image’s resolution with Bria while controlling color and transparency preservation.
- [Bria Product Dimensions](https://raywake.si/explore/bria/product-dimensions): Add supplied physical dimension callouts to a product photo; the model does not measure the product for you.
- [Background Remover](https://raywake.si/explore/bria/background/remove): Remove the background from an existing product or portrait photo.
- [Bria's VRMBG 3.0](https://raywake.si/explore/bria/video/background-removal/v3): Isolate a moving subject with Bria VRMBG 3.0.
- [Bytedance Omnihuman V1.5](https://raywake.si/explore/bytedance/omnihuman/v1.5): Animate a person in an image with a supplied audio recording.
- [Seedance 1 Pro Fast](https://raywake.si/explore/bytedance/seedance/v1/pro/fast/text-to-video): Create short motion studies with Seedance 1 Pro Fast.
- [Seedance 1.5 Pro](https://raywake.si/explore/bytedance/seedance/v1.5/pro/text-to-video): Generate short videos with optional audio using Seedance 1.5 Pro.
- [Seedream 4](https://raywake.si/explore/bytedance/seedream/v4/text-to-image): Create new images or edit references with Seedream 4.
- [Seedream 4.5](https://raywake.si/explore/seedream-4-5): Build a scene from a description. Explore architecture, landscapes and carefully composed visual stories.
- [Cartoonify](https://raywake.si/explore/cartoonify): Turn an existing photo into a cartoon treatment with Cartoonify.
- [Chatterbox](https://raywake.si/explore/chatterbox/text-to-speech): Turn your text into speech with Chatterbox and its voice controls.
- [Chatterboxhd](https://raywake.si/explore/resemble-ai/chatterboxhd/text-to-speech): Turn your text into speech with Chatterboxhd and its voice controls.
- [Chrono Edit](https://raywake.si/explore/chrono-edit): Apply a written change to a source image with Chrono Edit.
- [Chrono Edit LoRA](https://raywake.si/explore/chrono-edit-lora): Use the Chrono Edit LoRA variant for a source-image edit.
- [Chrono Edit Lora Gallery](https://raywake.si/explore/chrono-edit-lora-gallery/paintbrush): Turn sketched image regions into edits using Chrono Edit Paintbrush.
- [CogView](https://raywake.si/explore/cogview4): Generate images from detailed descriptions with CogView 4.
- [SDXL ControlNet Canny](https://raywake.si/explore/fast-sdxl-controlnet-canny): Guide an SDXL scene with edge structure using SDXL ControlNet Canny.
- [Cosmos 3 Super Image](https://raywake.si/explore/nvidia/cosmos-3-super/text-to-image): Create still images from text with Cosmos 3 Super Image.
- [Cosmos Predict 2.5](https://raywake.si/explore/cosmos-predict-2.5/text-to-video): Create text-guided video scenes with Cosmos Predict 2.5.
- [Cosmos Predict 2.5 Distilled](https://raywake.si/explore/cosmos-predict-2.5/distilled/text-to-video): Generate text-guided video with the distilled Cosmos Predict 2.5 model.
- [Creative Upscaler](https://raywake.si/explore/creative-upscaler): Enlarge and reinterpret image detail with Creative Upscaler.
- [DeepSeek Janus-Pro](https://raywake.si/explore/janus): Create an image from a text brief with DeepSeek Janus-Pro.
- [DreamOmni2](https://raywake.si/explore/dreamomni2/edit): Edit supplied images with a written instruction using DreamOmni2.
- [Dreamshaper](https://raywake.si/explore/dreamshaper): Explore illustrated or photographic concepts from text with Dreamshaper.
- [Eleven v4](https://raywake.si/explore/elevenlabs/tts/eleven-v4): Read a voiceover script with Eleven v4, using audio tags to direct delivery.
- [Eleven v4 Turbo](https://raywake.si/explore/elevenlabs/tts/eleven-v4-turbo): Prepare an alternate narration with Eleven v4 Turbo.
- [ElevenLabs Audio Isolation](https://raywake.si/explore/elevenlabs/audio-isolation): Supply a recording to isolate speech from background sound.
- [ElevenLabs Dubbing](https://raywake.si/explore/elevenlabs/dubbing): Dub an existing audio recording into a selected language with ElevenLabs.
- [ElevenLabs TTS Multilingual v2](https://raywake.si/explore/elevenlabs/tts/multilingual-v2): Read narration in a supported language with ElevenLabs Multilingual v2.
- [ElevenLabs TTS Turbo v2.5](https://raywake.si/explore/elevenlabs/tts/turbo-v2.5): Adjust a spoken script’s pace with ElevenLabs Turbo v2.5.
- [ElevenLabs Voice Changer](https://raywake.si/explore/elevenlabs/voice-changer): Transform a recorded performance with ElevenLabs Voice Changer.
- [ElevenLabs Music](https://raywake.si/explore/elevenlabs-music): Describe the mood, instruments and direction of a track, then create music for your next idea.
- [Elevenlabs Sound Effects V2](https://raywake.si/explore/elevenlabs/sound-effects/v2): Design a short sound effect from a description with ElevenLabs Sound Effects V2.
- [Elevenlabs Tts Eleven V3](https://raywake.si/explore/elevenlabs/tts/eleven-v3): Create a spoken version of a script with Eleven v3.
- [Emu 3.5 Image](https://raywake.si/explore/emu-3.5-image/text-to-image): Create images from scene descriptions with Emu 3.5 Image.
- [Ernie Image](https://raywake.si/explore/ernie-image): Generate images from text with Baidu’s Ernie Image and optional prompt expansion.
- [Ernie Image Turbo](https://raywake.si/explore/ernie-image/turbo): Ernie Image Turbo generates text-based images with a shorter sampling setup.
- [Bria Extract Object](https://raywake.si/explore/bria/extract-object): Select a particular object from a photograph with a written description.
- [FLUX.2 [dev]](https://raywake.si/explore/flux-2): Build a new scene from text with FLUX.2 [dev].
- [FLUX.2 [flash]](https://raywake.si/explore/flux-2/flash): Explore text-based image variations with FLUX.2 [flash].
- [FLUX 2 Lora Gallery Hdr Style](https://raywake.si/explore/flux-2-lora-gallery/hdr-style): Create a dramatic tonal treatment from text with FLUX 2 HDR Style.
- [FLUX 2 Lora Gallery Realism](https://raywake.si/explore/flux-2-lora-gallery/realism): Generate a photographic-style concept with FLUX 2 Realism.
- [FLUX 2 Turbo](https://raywake.si/explore/flux-2/turbo): Generate a fresh image from a scene brief with FLUX 2 Turbo.
- [FLUX.3 Image to Video](https://raywake.si/explore/blackforestlabs/flux-3/image-to-video): Turn one still image into a moving video with FLUX 3.
- [FLUX.1 Kontext [max]](https://raywake.si/explore/flux-pro/kontext/max/text-to-image): Explore an alternative text-to-image composition with Kontext Max.
- [FLUX.1 Kontext [pro]](https://raywake.si/explore/flux-pro/kontext/text-to-image): Create a new image from a scene description with Kontext Pro.
- [FLUX.1 Krea [dev]](https://raywake.si/explore/flux-1/krea): Explore a photographic scene from text with FLUX.1 Krea [dev].
- [FLUX.1 Krea [dev]](https://raywake.si/explore/flux/krea): FLUX.1 Krea [dev] generates images from detailed text descriptions.
- [FLUX.1 Krea [dev] Redux](https://raywake.si/explore/flux-1/krea/redux): Reinterpret a source image with FLUX.1 Krea [dev] Redux.
- [FLUX.1 SRPO [dev]](https://raywake.si/explore/flux-1/srpo): Build a detailed visual brief for FLUX.1 SRPO [dev].
- [FLUX.1 SRPO [dev]](https://raywake.si/explore/flux/srpo): FLUX.1 SRPO [dev] generates images from text with an emphasis on visual aesthetics.
- [FLUX.1 [dev]](https://raywake.si/explore/flux-1/dev): Develop a text-to-image concept with FLUX.1 [dev].
- [FLUX.1 [dev]](https://raywake.si/explore/flux/dev): Generate text-based images with FLUX.1 [dev], or use Redux for image-led variations.
- [FLUX.1 [dev] Redux](https://raywake.si/explore/flux-1/dev/redux): Generate a new interpretation of an image with FLUX.1 [dev] Redux.
- [FLUX.1 Differential Diffusion](https://raywake.si/explore/flux-general/differential-diffusion): Use FLUX.1 differential diffusion to edit an image with a change map.
- [FLUX.1 RF-Inversion](https://raywake.si/explore/flux-general/rf-inversion): Edit a source image through FLUX.1 RF-Inversion using a text instruction.
- [FLUX.1 [schnell]](https://raywake.si/explore/flux-1/schnell): Sketch a scene from text with FLUX.1 [schnell].
- [FLUX.1 [schnell]](https://raywake.si/explore/flux/schnell): FLUX.1 [schnell] creates text-based images with a short sampling schedule.
- [FLUX.1 [schnell] Redux](https://raywake.si/explore/flux-1/schnell/redux): Explore image-led variations with FLUX.1 [schnell] Redux.
- [FLUX.2 [klein] 4B](https://raywake.si/explore/flux-2/klein/4b): Draft a text-led composition with FLUX.2 [klein] 4B.
- [FLUX.2 [klein] 4B Base](https://raywake.si/explore/flux-2/klein/4b/base): Control a text-generated scene with FLUX.2 [klein] 4B Base.
- [FLUX.2 [klein] 9B](https://raywake.si/explore/flux-2/klein/9b): Create a scene from a focused brief with FLUX.2 [klein] 9B.
- [FLUX.2 [klein] 9B Base](https://raywake.si/explore/flux-2/klein/9b/base): Develop a scene with exclusions using FLUX.2 [klein] 9B Base.
- [FLUX1.1 [pro]](https://raywake.si/explore/flux-pro/v1.1): Create a new image from a written brief with FLUX1.1 [pro].
- [FLUX1.1 [pro] ultra](https://raywake.si/explore/flux-pro/v1.1-ultra): Explore a scene with optional image guidance using FLUX1.1 [pro] ultra.
- [Fabric 1.0](https://raywake.si/explore/veed/fabric-1.0): Combine an image and audio with VEED Fabric 1.0 to make a talking video.
- [Face Retoucher](https://raywake.si/explore/retoucher): Retouch a face photo with Face Retoucher.
- [Feynobg Background Remover](https://raywake.si/explore/feynobg): Extract a foreground with Feynobg Background Remover.
- [Fibo](https://raywake.si/explore/bria/fibo/generate): Generate a composed image with Fibo, using a written or structured prompt.
- [Fibo Edit [Add Object by Text]](https://raywake.si/explore/bria/fibo-edit/add_object_by_text): Use Fibo Edit [Add Object by Text] to add an object at a described position.
- [Fibo Edit [Colorize]](https://raywake.si/explore/bria/fibo-edit/colorize): Use Fibo Edit [Colorize] to change a photo’s color treatment.
- [Fibo Edit [Erase by Text]](https://raywake.si/explore/bria/fibo-edit/erase_by_text): Use Fibo Edit [Erase by Text] to remove a named object from a photo.
- [Fibo Edit [Relight]](https://raywake.si/explore/bria/fibo-edit/relight): Use Fibo Edit [Relight] to change the direction and character of photo lighting.
- [Fibo Edit [Replace Object by Text]](https://raywake.si/explore/bria/fibo-edit/replace_object_by_text): Use Fibo Edit [Replace Object by Text] to replace an object using a written instruction.
- [Fibo Edit [Reseason]](https://raywake.si/explore/bria/fibo-edit/reseason): Use Fibo Edit [Reseason] to change the season in an existing scene.
- [Fibo Edit [Restore]](https://raywake.si/explore/bria/fibo-edit/restore): Use Fibo Edit [Restore] to restore a damaged or blurred photo.
- [Fibo Edit [Restyle]](https://raywake.si/explore/bria/fibo-edit/restyle): Use Fibo Edit [Restyle] to redraw a photo in a chosen artistic style.
- [Fibo Edit [Rewrite Text]](https://raywake.si/explore/bria/fibo-edit/rewrite_text): Use Fibo Edit [Rewrite Text] to replace lettering inside an image.
- [Fibo Edit [Sketch to Image]](https://raywake.si/explore/bria/fibo-edit/sketch_to_colored_image): Use Fibo Edit [Sketch to Image] to turn a sketch into a colored image.
- [Fibo Gen 1.5 Text to Image](https://raywake.si/explore/bria/fibo-gen-1.5/text-to-image): Generate a composed image with Fibo Gen 1.5 Text to Image, using a written or structured prompt.
- [Fibo Lite](https://raywake.si/explore/bria/fibo-lite/generate): Generate a composed image with Fibo Lite, using a written or structured prompt.
- [Finegrain Eraser](https://raywake.si/explore/finegrain-eraser): Remove a described object from a photo with Finegrain Eraser.
- [Florence-2 Phrase Grounding](https://raywake.si/explore/florence-2-large/caption-to-phrase-grounding): Locate phrases from an image caption with Florence-2 phrase grounding.
- [Florence-2 Dense Region Captions](https://raywake.si/explore/florence-2-large/dense-region-caption): Describe individual image regions with Florence-2 dense region captions.
- [Florence-2 Object Detection](https://raywake.si/explore/florence-2-large/object-detection): Locate objects in a source image with Florence-2 object detection.
- [Florence-2 OCR with Regions](https://raywake.si/explore/florence-2-large/ocr-with-region): Read text from an image together with its detected regions.
- [Florence-2 Open-Vocabulary Detection](https://raywake.si/explore/florence-2-large/open-vocabulary-detection): Search an image for a named object with Florence-2 open-vocabulary detection.
- [Florence-2 Referring-Expression Segmentation](https://raywake.si/explore/florence-2-large/referring-expression-segmentation): Select an image region using a referring expression with Florence-2 segmentation.
- [Florence-2 Region Proposals](https://raywake.si/explore/florence-2-large/region-proposal): Find candidate regions in an image with Florence-2 region proposals.
- [Flow-Edit](https://raywake.si/explore/flowedit): Transform a photo using source and target descriptions with Flow-Edit.
- [Flux 2 Flex](https://raywake.si/explore/flux-2-flex): Text-to-image generation with FLUX.2 [flex] from Black Forest Labs. Features adjustable inference steps and guidance scale for fine-tuned control. Enhanced typography and text rendering capabilities.
- [FLUX 2 Ballpoint Pen Sketch](https://raywake.si/explore/flux-2-lora-gallery/ballpoint-pen-sketch): Generate a ballpoint sketch from text with FLUX 2 Ballpoint Pen Sketch.
- [FLUX 2 Digital Comic Art](https://raywake.si/explore/flux-2-lora-gallery/digital-comic-art): Build a comic-style scene from text with FLUX 2 Digital Comic Art.
- [Flux 2 Lora Gallery](https://raywake.si/explore/flux-2-lora-gallery/multiple-angles): Generates same object from different angles (azimuth/elevation)
- [FLUX 2 Satellite View Style](https://raywake.si/explore/flux-2-lora-gallery/satellite-view-style): Create an imagined aerial scene with FLUX 2 Satellite View Style.
- [FLUX 2 Sepia Vintage](https://raywake.si/explore/flux-2-lora-gallery/sepia-vintage): Generate a sepia-toned photographic concept with FLUX 2 Sepia Vintage.
- [Flux 2 Lora Gallery](https://raywake.si/explore/flux-2-lora-gallery/virtual-tryon): Virtual clothing try-on (2 images: person + garment)
- [Flux 2 Max](https://raywake.si/explore/flux-2-max): FLUX.2 [max] delivers state-of-the-art image generation and advanced image editing with exceptional realism, precision, and consistency.
- [Flux 2 Max](https://raywake.si/explore/flux-2-max/edit): FLUX.2 Max edits supplied images using a text instruction.
- [FLUX.2 [pro]](https://raywake.si/explore/flux-2-pro): Shape a product shot, an editorial scene or a new visual idea from a written prompt.
- [FLUX.3 Video](https://raywake.si/explore/blackforestlabs/flux-3/edit-video): FLUX.3 Edit Video [FAST] is Black Forest Labs' frontier video model. This endpoint edits an existing video from natural-language instructions, applying targeted changes while preserving the rest of the scene.
- [FLUX.3 Image](https://raywake.si/explore/blackforestlabs/flux-3/text-to-image): Create an image from text or edit with up to ten ordered reference images using FLUX.3 Image.
- [Fooocus Image-to-Image](https://raywake.si/explore/fast-fooocus-sdxl/image-to-image): Rework an existing image with Fooocus using a text instruction.
- [Fooocus](https://raywake.si/explore/fooocus): Generate styled images, edit a masked region or upscale and vary an existing image with Fooocus.
- [GLM Image](https://raywake.si/explore/glm-image): Generate illustrated scenes and text-bearing visuals with GLM Image.
- [GPT Image 1 Mini](https://raywake.si/explore/gpt-image-1-mini): Generate images from written briefs with GPT Image 1 Mini.
- [GPT Image 2](https://raywake.si/explore/openai/gpt-image-2): Create GPT Image 2 visuals from text or switch to image editing.
- [GPT Image 2.5 Flare](https://raywake.si/explore/openai/gpt-image-2.5/flare/text-to-image): Generate GPT Image 2.5 Flare visuals, with a separate editing operation.
- [GPT Image 2.5](https://raywake.si/explore/gpt-image-2-5): Turn a detailed brief into an image, then explore a different interpretation with image editing.
- [GPT Image 1.5](https://raywake.si/explore/gpt-image-1.5): Create detailed compositions from text with GPT Image 1.5.
- [Gemini 2.5 Flash Image](https://raywake.si/explore/gemini-25-flash-image): Google's famous original image generation and editing model, a.k.a Nano Banana
- [Gemini 2.5 Flash Image](https://raywake.si/explore/gemini-25-flash-image/edit): Edit supplied images with Gemini 2.5 Flash Image and a text instruction.
- [Gemini 3 Pro Image](https://raywake.si/explore/gemini-3-pro-image-preview): Create images from text with Gemini 3 Pro Image, with a separate image editing mode.
- [Gemini 3.1 Flash Image](https://raywake.si/explore/gemini-3.1-flash-image-preview): Generate images with Gemini 3.1 Flash Image, including very wide or tall layouts.
- [Gemini 3.1 Flash Tts](https://raywake.si/explore/gemini-3.1-flash-tts): Convert written text to expressive speech with Gemini 3.1 Flash TTS.
- [Gemini 3.8 Flash Lite TTS](https://raywake.si/explore/google/gemini-3.8-flash-lite-tts): Use Flash Lite to read a short tutorial script or alternate between two configured speakers.
- [Gemini 3.8 Flash TTS](https://raywake.si/explore/google/gemini-3.8-flash-tts): Prepare narration or a two-person conversation from a written script.
- [Gemini Omni Flash](https://raywake.si/explore/google/gemini-omni-flash/image-to-video): Animate a still image into a short video with Gemini Omni Flash.
- [Gemini Omni Flash 1.1](https://raywake.si/explore/google/gemini-omni-flash/v1.1/text-to-video): Create short videos with synchronized audio using Gemini Omni Flash 1.1.
- [Ghiblify Images](https://raywake.si/explore/ghiblify): Explore an animated illustration treatment of a photo with Ghiblify Images.
- [Google Veo 3](https://raywake.si/explore/veo-3): Generate a short video from text with Google Veo 3, optionally including audio.
- [Google Veo 3 Fast](https://raywake.si/explore/veo-3-fast): Create text-led short videos with the Google Veo 3 Fast operation.
- [Google Veo 3 Fast Image to Video](https://raywake.si/explore/veo-3-fast-image-to-video): Animate a source image with Google Veo 3 Fast, with optional generated audio.
- [Google Veo 3 Image to Video](https://raywake.si/explore/veo-3-image-to-video): Google Veo 3 animates one source image from a motion prompt, with optional generated audio.
- [Google Virtual Try On](https://raywake.si/explore/google/virtual-try-on): Preview clothing on a person with Google Virtual Try On.
- [Grok Imagine Image](https://raywake.si/explore/xai/grok-imagine-image): Grok Imagine Image creates images from text and edits references through a separate mode.
- [Grok Imagine Image Quality](https://raywake.si/explore/xai/grok-imagine-image/quality/text-to-image): Grok Imagine Image Quality is a dedicated text-to-image operation.
- [Grok Imagine Image 2.0](https://raywake.si/explore/xai/grok-imagine-image/v2.0/text-to-image): Grok Imagine Image 2.0 generates images from text and offers a separate reference editing mode.
- [Grok Imagine Video](https://raywake.si/explore/xai/grok-imagine-video/text-to-video): Grok Imagine Video creates video with audio from text or animates an image in a separate mode.
- [Grok Imagine Video 1.5](https://raywake.si/explore/xai/grok-imagine-video/v1.5/text-to-video): Grok Imagine Video 1.5 creates video with audio and offers a separate image animation mode.
- [MiniMax H3 Max Styles](https://raywake.si/explore/minimax/h3-max/styles/16bit-pixel): Create a video in H3 Max’s 16-bit pixel style.
- [MiniMax H3 Max Recast](https://raywake.si/explore/minimax/h3-max/recast): Replace people in a video with reference photos using MiniMax H3 Max Recast.
- [MiniMax H3 Max Turbo](https://raywake.si/explore/minimax/h3-max-turbo/text-to-video): Generate a video scene with MiniMax H3 Max Turbo.
- [Happy Horse](https://raywake.si/explore/alibaba/happy-horse/text-to-video): Create short videos with synchronized sound from a written scene.
- [Happy Horse 1.1](https://raywake.si/explore/alibaba/happy-horse/v1.1/text-to-video): Generate short videos with native audio using Happy Horse 1.1.
- [Hidream O1 Image](https://raywake.si/explore/hidream-o1-image): Generate images from text with the full HiDream O1 Image model.
- [Hidream O1 Image](https://raywake.si/explore/hidream-o1-image/dev): Try text-based image concepts with HiDream O1 Image Dev.
- [Hotel Lobby Studio Video](https://raywake.si/explore/hotel-lobby-studio-video): Place one or two people into a prepared dance clip with Hotel Lobby Studio Video.
- [Hunyuan Video (Video-to-Video)](https://raywake.si/explore/hunyuan-video/video-to-video): Reinterpret an existing clip with Hunyuan Video video-to-video.
- [Hunyuan World](https://raywake.si/explore/hunyuan_world): Expand an image into a panorama with Hunyuan World.
- [Ideogram Custom Models](https://raywake.si/explore/ideogram/custom-models/generate): Generate images with a trained Ideogram custom model.
- [Ideogram V3 Layerize Text](https://raywake.si/explore/ideogram/v3/layerize-text): Ideogram Layerize takes an existing flat graphic, removes text, and returns structured text containers you can edit/recompose in html or json format.
- [Ideogram](https://raywake.si/explore/ideogram/v3/reframe): Reframe an existing image into a new canvas with Ideogram V3.
- [Ideogram Object Removal](https://raywake.si/explore/ideogram/object-removal): Remove a selected object from an existing image using a mask.
- [Ideogram Remove Background](https://raywake.si/explore/ideogram/remove-background): Prepare a foreground cutout from a product or portrait photo.
- [Ideogram V3](https://raywake.si/explore/ideogram/v3): Generate typography and style-guided visuals with Ideogram V3.
- [Ideogram Upscale](https://raywake.si/explore/ideogram/upscale): Enlarge an existing image while reviewing the balance between source resemblance and added detail.
- [Ideogram V2](https://raywake.si/explore/ideogram/v2): Create text-bearing graphics with Ideogram V2.
- [Ideogram V2 Turbo](https://raywake.si/explore/ideogram/v2/turbo): Draft text-based graphics with Ideogram V2 Turbo.
- [Ideogram V2A](https://raywake.si/explore/ideogram/v2a): Build image and typography concepts with Ideogram V2A.
- [Ideogram V2A Turbo](https://raywake.si/explore/ideogram/v2a/turbo): Try poster and illustration drafts with Ideogram V2A Turbo.
- [Ideogram Character](https://raywake.si/explore/ideogram/character): Place a reference character into a described scene with Ideogram Character.
- [Ideogram V4.5 Edit](https://raywake.si/explore/ideogram/v4.5/edit): Revise an existing image with Ideogram 4.5 Edit.
- [Illusion Diffusion](https://raywake.si/explore/illusion-diffusion): Build a scene around a reference pattern with Illusion Diffusion.
- [Image Editing Tools](https://raywake.si/explore/image-editing/age-progression): Explore apparent aging through the default Age Progression mode of Image Editing Tools.
- [Depth Anything V2 Preprocessor](https://raywake.si/explore/image-preprocessors/depth-anything/v2): Produce a depth-map image from a photograph with Depth Anything V2.
- [HED Edge Preprocessor](https://raywake.si/explore/image-preprocessors/hed): Extract a line map from an image with the HED detector.
- [Lineart Preprocessor](https://raywake.si/explore/image-preprocessors/lineart): Create a line-art map from a source image.
- [MLSD Line Preprocessor](https://raywake.si/explore/image-preprocessors/mlsd): Detect line segments in an image with MLSD.
- [PiDi Edge Preprocessor](https://raywake.si/explore/image-preprocessors/pidi): Extract contours with the PiDi image preprocessor.
- [SAM Segmentation Preprocessor](https://raywake.si/explore/image-preprocessors/sam): Produce a segmentation-map image with the SAM preprocessor.
- [Scribble Preprocessor](https://raywake.si/explore/image-preprocessors/scribble): Create a scribble-style guide from an existing image.
- [TeeD Edge Preprocessor](https://raywake.si/explore/image-preprocessors/teed): Generate a TeeD line map for a source photograph.
- [Zoe Depth Preprocessor](https://raywake.si/explore/image-preprocessors/zoe): Estimate an image’s depth structure with the Zoe preprocessor.
- [Image2Pixel](https://raywake.si/explore/image2pixel): Clean up a pixel-art image with Image2Pixel.
- [Infinity Star](https://raywake.si/explore/infinity-star/text-to-video): Create short scene concepts from text with Infinity Star.
- [Inpainting sdxl and sd](https://raywake.si/explore/inpaint): Replace a masked part of an image with the Inpainting route.
- [Juggernaut Flux Base](https://raywake.si/explore/rundiffusion/juggernaut-flux/base): Generate an image from a scene description with Juggernaut Flux Base.
- [Juggernaut Flux Lightning](https://raywake.si/explore/rundiffusion/juggernaut-flux/lightning): Generate images from a written scene with Juggernaut Flux Lightning.
- [Juggernaut Flux Pro](https://raywake.si/explore/rundiffusion/juggernaut-flux/pro): Create text-led images with Juggernaut Flux Pro and adjustable prompt guidance.
- [Kling 1.0](https://raywake.si/explore/kling-video/v1/standard/effects): Generate video clips from your prompts using Kling 1.0
- [Kling 1.5](https://raywake.si/explore/kling-video/v1.5/pro/effects): Generate video clips from your prompts using Kling 1.5 (pro)
- [Kling 1.6](https://raywake.si/explore/kling-video/v1.6/pro/effects): Generate video clips from your prompts using Kling 1.6 (pro)
- [Kling 1.6](https://raywake.si/explore/kling-video/v1.6/standard/effects): Generate video clips from your prompts using Kling 1.6 (std)
- [Kling AI Avatar](https://raywake.si/explore/kling-video/v1/standard/ai-avatar): Kling AI Avatar Standard:  Endpoint for creating avatar videos with realistic humans, animals, cartoons, or stylized characters
- [Kling AI Avatar Pro](https://raywake.si/explore/kling-video/v1/pro/ai-avatar): Kling AI Avatar Pro: The premium endpoint for creating avatar videos with realistic humans, animals, cartoons, or stylized characters
- [Kling AI Avatar v2](https://raywake.si/explore/kling-video/ai-avatar/v2/standard): Animate an avatar image using a supplied audio recording.
- [Kling O3 Image](https://raywake.si/explore/kling-image/o3/text-to-image): Generate single images or a related image series with Kling O3 Image.
- [Kling V3 Image](https://raywake.si/explore/kling-image/v3/text-to-image): Create images from a scene description with Kling Image V3.
- [Kling O1 Edit Video [Pro]](https://raywake.si/explore/kling-video/o1/video-to-video/edit): Edit a source video through the Kling O1 Pro operation.
- [Kling O1 Edit Video [Standard]](https://raywake.si/explore/kling-video/o1/standard/video-to-video/edit): Revise an existing clip with Kling O1 Standard video editing.
- [Kling O1](https://raywake.si/explore/kling-video/o1/image-to-video): Generate a video by taking a start frame and an end frame, animating the transition between them while following text-driven style and scene guidance.
- [Kling O1 Standard](https://raywake.si/explore/kling-video/o1/standard/image-to-video): Generate a video by taking a start frame and an end frame, animating the transition between them while following text-driven style and scene guidance.
- [Kling O1 Image](https://raywake.si/explore/kling-image/o1): Edit and combine image references with Kling O1 Image.
- [Kling O3 Pro](https://raywake.si/explore/kling-video/o3/pro/text-to-video): Generate and edit videos with Kling O3 Pro, using text, images, characters or an existing clip.
- [Kling O3 Standard](https://raywake.si/explore/kling-video/o3/standard/text-to-video): Generate and edit videos with Kling O3 Standard, using text, images, characters or an existing clip.
- [Kling TTS](https://raywake.si/explore/kling-video/v1/tts): Generate speech from text prompts and different voices using the Kling TTS model, which leverages advanced AI techniques to create high-quality text-to-speech.
- [Kling O3 4K](https://raywake.si/explore/kling-video/o3/4k/text-to-video): Generate and edit videos with Kling O3 4K, using text, images, characters or an existing clip.
- [Kling 2.5 Turbo Standard](https://raywake.si/explore/kling-video/v2.5-turbo/standard/image-to-video): Turn a source image into a short clip with Kling V2.5 Turbo Standard.
- [Kling Video to Audio](https://raywake.si/explore/kling-video/video-to-audio): Generate a soundtrack for an existing video with Kling Video to Audio.
- [Kling 3 Turbo Standard](https://raywake.si/explore/kling-video/v3/turbo/standard/text-to-video): Generate a text-described clip with Kling V3 Turbo Standard.
- [Kling 3 4K](https://raywake.si/explore/kling-video/v3/4k/text-to-video): Create text-driven clips with the Kling V3 4K operation.
- [Kling 3 Turbo Pro](https://raywake.si/explore/kling-video/v3/turbo/pro/text-to-video): Create a scene from text with Kling V3 Turbo Pro.
- [Kling Video v2.6 Motion Control [Standard]](https://raywake.si/explore/kling-video/v2.6/standard/motion-control): Apply a reference performance to a character image with Kling V2.6 Standard Motion Control.
- [Kling 2.6 Pro](https://raywake.si/explore/kling-video/v2.6/pro/text-to-video): Generate short scenes with optional audio using Kling V2.6 Pro.
- [Kling 3 Pro](https://raywake.si/explore/kling-3-pro): Give a still image a sense of motion. Direct the scene with a prompt and turn a frame into a short film.
- [Kling 3 Standard](https://raywake.si/explore/kling-video/v3/standard/text-to-video): Create a scene from text with Kling V3 Standard, or choose a visual source operation.
- [Kling 2.5 Turbo Pro](https://raywake.si/explore/kling-video/v2.5-turbo/pro/text-to-video): Create a 5- or 10-second scene with Kling V2.5 Turbo Pro.
- [Kokoro TTS](https://raywake.si/explore/kokoro/american-english): Read an English script aloud with Kokoro American English.
- [Kolors](https://raywake.si/explore/kolors): Create images from a visual brief with Kolors.
- [Krea 2 Large](https://raywake.si/explore/krea/v2/large/text-to-image): Generate an image from a written brief with Krea 2 Large.
- [Krea 2 Medium](https://raywake.si/explore/krea/v2/medium/text-to-image): Create a text-described image with Krea 2 Medium.
- [Krea 2 Medium Turbo](https://raywake.si/explore/krea/v2/medium/turbo/text-to-image): Create images with the Krea 2 Medium Turbo partner operation.
- [Krea 2 Turbo](https://raywake.si/explore/krea-2/turbo): Create images from text with the fal Krea 2 Turbo operation.
- [LTX 2.3 Fast](https://raywake.si/explore/ltx-2.3/text-to-video/fast): Plan a continuous shot from a scene description with LTX 2.3 Fast.
- [LTX 2.3 Pro](https://raywake.si/explore/ltx-2.3/text-to-video): Generate a scene with LTX 2.3 Pro’s duration and framing presets.
- [LTX-2.3 22B](https://raywake.si/explore/ltx-2.3-22b/text-to-video): Create video with optional sound using LTX-2.3 22B’s sampling controls.
- [LTX-2.3 22B Distilled](https://raywake.si/explore/ltx-2.3-22b/distilled/text-to-video): Generate video and optional audio with the distilled LTX-2.3 22B route.
- [LCM Diffusion](https://raywake.si/explore/fast-lcm-diffusion): LCM Diffusion generates images from text with a low-step sampling approach.
- [Live Portrait](https://raywake.si/explore/live-portrait): Transfer a driving video’s facial motion to a still portrait.
- [Longcat Image](https://raywake.si/explore/longcat-image): Generate still images from a written scene with Longcat Image.
- [LTX-2.3 Quality](https://raywake.si/explore/ltx-2.3-quality/text-to-video): Generate video and optional audio with the LTX-2.3 Quality route.
- [Luma Ray 3.2](https://raywake.si/explore/luma/agent/ray/v3.2/text-to-video): Generate a shot from text with Luma Ray 3.2.
- [Luma Uni-1](https://raywake.si/explore/luma/agent/uni-1/v1/text-to-image): Create still images from a scene description with Luma Uni-1.
- [Luma Uni-1 Max](https://raywake.si/explore/luma/agent/uni-1/v1/max): Generate a still with the Max tier of Luma Uni-1.
- [Lumina Image 2](https://raywake.si/explore/lumina-image/v2): Generate still images with Lumina Image 2 and guidance controls.
- [Lyria 3 Pro](https://raywake.si/explore/lyria3/pro): Create music from a written brief with Lyria 3 Pro.
- [Lyria 3.5](https://raywake.si/explore/google/lyria-3.5): Compose music from a written brief with Lyria 3.5.
- [Lyria3](https://raywake.si/explore/lyria3): Lyria 3 is most recent music model from Google
- [Marey Realism 1.5](https://raywake.si/explore/moonvalley/marey/t2v): Generate Marey video from text, with image and transfer operations available separately.
- [Marigold Depth Estimation](https://raywake.si/explore/imageutils/marigold-depth): Estimate a depth map from a single photo with Marigold.
- [Meta Muse Image](https://raywake.si/explore/meta/muse-image/text-to-image): Generate images from detailed instructions with Meta Muse Image.
- [Midas Depth Estimation](https://raywake.si/explore/imageutils/depth): Extract a depth representation from a photo with MiDaS.
- [MiniMax (Hailuo AI) Music](https://raywake.si/explore/minimax-music): Generate music from text prompts using the MiniMax model, which leverages advanced AI techniques to create high-quality, diverse musical compositions.
- [MiniMax (Hailuo AI) Music v1.5](https://raywake.si/explore/minimax-music/v1.5): Generate music from text prompts using the MiniMax model, which leverages advanced AI techniques to create high-quality, diverse musical compositions.
- [MiniMax Image-01](https://raywake.si/explore/minimax/image-01): Generate images from text, with a separate subject-reference operation.
- [MiniMax Video-01](https://raywake.si/explore/minimax/video-01): Generate video clips from your prompts using MiniMax model
- [MiniMax Video-01 Director](https://raywake.si/explore/minimax/video-01-director): Generate video clips more accurately with respect to natural language descriptions and using camera movement instructions for shot control.
- [MiniMax Video-01 Live](https://raywake.si/explore/minimax/video-01-live): Generate video clips from your prompts using MiniMax model
- [MiniMax (Hailuo AI) Video 01 Subject Reference](https://raywake.si/explore/minimax/video-01-subject-reference): Generate video clips maintaining consistent, realistic facial features and identity across dynamic video content
- [MiniMax H3 Max](https://raywake.si/explore/minimax/h3-max/text-to-video): H3 Max is a post-trained variant of MiniMax H3, tuned for stronger prompt adherence and better aesthetics while co-optimized with our custom inference stack for higher throughput with no compromises on output quality
- [MiniMax H3](https://raywake.si/explore/minimax/h3/text-to-video): Generate video from a scene description with MiniMax H3.
- [MiniMax H3 LoRA](https://raywake.si/explore/minimax/h3/image-to-video/lora): References into video with synchronized audio using MiniMax H3
- [MiniMax Hailuo 02 Pro](https://raywake.si/explore/minimax/hailuo-02/pro/text-to-video): Create a text-driven video scene with MiniMax Hailuo 02 Pro.
- [MiniMax Hailuo 02 Standard](https://raywake.si/explore/minimax/hailuo-02/standard/text-to-video): Generate a shot with MiniMax Hailuo 02 Standard.
- [MiniMax Hailuo 2.3 Fast [Pro] (Image to Video)](https://raywake.si/explore/minimax/hailuo-2.3-fast/pro/image-to-video): MiniMax Hailuo-2.3-Fast Image To Video API (Pro, 1080p): Advanced fast image-to-video generation model with 1080p resolution
- [MiniMax Hailuo 2.3 Fast](https://raywake.si/explore/minimax/hailuo-2.3-fast/standard/image-to-video): Animate a still with MiniMax Hailuo 2.3 Fast Standard.
- [MiniMax Hailuo 2.3 Pro](https://raywake.si/explore/minimax/hailuo-2.3/pro/text-to-video): Create a Hailuo 2.3 Pro video from a written scene description.
- [MiniMax Hailuo 2.3 Standard](https://raywake.si/explore/minimax/hailuo-2.3/standard/text-to-video): Turn a scene brief into a Hailuo 2.3 Standard video.
- [MiniMax Music 3](https://raywake.si/explore/minimax/music-3): MiniMax Music 3 is a high-performance music generation model for creating complete songs up to five minutes long
- [MiniMax Speech 2.8 [HD]](https://raywake.si/explore/minimax/speech-2.8-hd): Produce spoken narration with MiniMax Speech 2.8 HD.
- [MiniMax Speech-02 HD](https://raywake.si/explore/minimax/speech-02-hd): Convert written text to speech with MiniMax Speech-02 HD.
- [MiniMax Speech-02 Turbo](https://raywake.si/explore/minimax/speech-02-turbo): Turn your text into speech with MiniMax Speech-02 Turbo and its voice controls.
- [MiniMax Voice Cloning](https://raywake.si/explore/minimax/voice-clone): Create a voice reference from sample audio, with optional preview speech.
- [MiniMax Hailuo 02 Fast](https://raywake.si/explore/minimax/hailuo-02-fast/image-to-video): Animate a starting image with MiniMax Hailuo 02 Fast.
- [Minimax Music](https://raywake.si/explore/minimax-music/v2): Generate music from text prompts using the MiniMax Music 2.0 model, which leverages advanced AI techniques to create high-quality, diverse musical compositions.
- [Minimax Music 2.5](https://raywake.si/explore/minimax-music/v2.5): MiniMax Music 2.5 creates complete tracks with singing, backing music, and detailed arrangements from lyrics and a style description.
- [Minimax Music 2.6](https://raywake.si/explore/minimax-music/v2.6): Generate songs or instrumental music with MiniMax Music 2.6.
- [Moondream 2 Object Detection](https://raywake.si/explore/moondream2/object-detection): Locate a named object in a photo with Moondream 2 Object Detection.
- [MuseTalk](https://raywake.si/explore/musetalk): Synchronize a face in a source video with supplied speech audio.
- [Nano Banana](https://raywake.si/explore/nano-banana): Create Nano Banana images from text or choose its separate editing operation.
- [Nano Banana 2](https://raywake.si/explore/nano-banana-2): Generate or edit Nano Banana 2 images with selectable resolution and framing.
- [Nano Banana 2 Lite](https://raywake.si/explore/google/nano-banana-2-lite): Create image drafts with Nano Banana 2 Lite.
- [Nano Banana Pro](https://raywake.si/explore/nano-banana-pro): Explore portraits, compositions and visual ideas. Start with words or bring an image into the studio to edit.
- [Object Removal](https://raywake.si/explore/object-removal): Remove a named object using the default Object Removal operation.
- [Omni Zero](https://raywake.si/explore/omni-zero): Any pose, any style, any identity
- [Omnigen V2](https://raywake.si/explore/omnigen-v2): Generate or edit images using a prompt and optional image references.
- [Optimized Latent Consistency (SDv1.5)](https://raywake.si/explore/lcm-sd15-i2i): Rework an existing image with the SDv1.5 Latent Consistency model.
- [Ovis Image](https://raywake.si/explore/ovis-image): Generate Ovis Image illustrations and text-bearing visuals from a written brief.
- [P-Video 2 Pro](https://raywake.si/explore/p-video-2-pro): Create P-Video 2 Pro clips from text or optional first and last frame images.
- [Phota Enhance](https://raywake.si/explore/phota/enhance): Enhance a supplied image with optional Phota identity profiles.
- [Phota](https://raywake.si/explore/phota): Create Phota images from a brief or edit existing images in its alternate operation.
- [Photo Restoration](https://raywake.si/explore/image-apps-v2/photo-restoration): Restore old or damaged photos by fixing colors, scratches, and resolution.
- [PhotoMaker](https://raywake.si/explore/photomaker): Create personalized portraits from an archive of reference images and a prompt.
- [Pika 2.2](https://raywake.si/explore/pika/v2.2/text-to-video): Generate Pika 2.2 video from text, images or combined scene references.
- [PixVerse C1](https://raywake.si/explore/pixverse/c1/text-to-video): Create PixVerse C1 video with optional audio or choose image and transition operations.
- [PixVerse Extend](https://raywake.si/explore/pixverse/extend): Continue an existing clip from its ending with PixVerse Extend.
- [PixVerse 3.5](https://raywake.si/explore/pixverse/v3.5/text-to-video): Turn a scene description into a clip with PixVerse 3.5.
- [PixVerse 3.5 Fast](https://raywake.si/explore/pixverse/v3.5/text-to-video/fast): Create a clip from text using PixVerse 3.5 Fast.
- [PixVerse 4](https://raywake.si/explore/pixverse/v4/text-to-video): Create a scene from text with PixVerse 4.
- [PixVerse 4 Fast](https://raywake.si/explore/pixverse/v4/text-to-video/fast): Produce a scene from a prompt with PixVerse 4 Fast.
- [PixVerse 4.5](https://raywake.si/explore/pixverse/v4.5/text-to-video): Generate a described scene with PixVerse 4.5.
- [PixVerse 4.5 Fast](https://raywake.si/explore/pixverse/v4.5/text-to-video/fast): Build a text-driven clip with PixVerse 4.5 Fast.
- [PixVerse 5](https://raywake.si/explore/pixverse/v5/text-to-video): Generate a scene from a text prompt with PixVerse 5.
- [PixVerse 5.5](https://raywake.si/explore/pixverse/v5.5/text-to-video): Generate a scene with optional audio using PixVerse 5.5.
- [PixVerse 5.6](https://raywake.si/explore/pixverse/v5.6/text-to-video): Make a text-driven video with optional sound in PixVerse 5.6.
- [PixVerse 6](https://raywake.si/explore/pixverse/v6/text-to-video): Create a 1–15-second clip with optional audio in PixVerse 6.
- [Pixelcut Background Remover](https://raywake.si/explore/pixelcut/background-removal): Prepare an image cutout with Pixelcut Background Remover.
- [Playground v2.5](https://raywake.si/explore/playground-v25): Create an image from a scene description with Playground v2.5.
- [Pony V7](https://raywake.si/explore/pony-v7): Generate a described image with Pony V7.
- [Post Processing](https://raywake.si/explore/post-processing/blur): Apply Gaussian or Kuwahara blur effects with adjustable radius and sigma parameters
- [Post Processing Sharpen](https://raywake.si/explore/post-processing/sharpen): Sharpen an image with basic, smart or contrast-adaptive sharpening.
- [Qwen Image 3 Image Editing](https://raywake.si/explore/alibaba/qwen-image-3/edit): Edit reference images with written instructions using Qwen Image 3.
- [Qwen Image 3.0](https://raywake.si/explore/qwen-image-3-0): Generate one image from text with Qwen Image 3.0 on Raywake.
- [Qwen Add Background](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/add-background): Place a clean-background subject in a new setting with Qwen Add Background.
- [Qwen Face to Full Portrait](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/face-to-full-portrait): Expand a close-up face into a portrait concept with Qwen Face to Full Portrait.
- [Qwen Group Photo](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/group-photo): Combine individual portraits into a shared scene with Qwen Group Photo.
- [Qwen Integrate Product](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/integrate-product): Blend a product into its surrounding scene with Qwen Integrate Product.
- [Qwen Lighting Restoration](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/lighting-restoration): Rework image lighting with Qwen Lighting Restoration.
- [Qwen Multiple Angles](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/multiple-angles): Explore a different viewpoint of a source scene with Qwen Multiple Angles.
- [Qwen Next Scene](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/next-scene): Plan a following storyboard frame with Qwen Next Scene.
- [Qwen Remove Element](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/remove-element): Erase a described scene element with Qwen Remove Element.
- [Qwen Remove Lighting](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/remove-lighting): Neutralize lighting effects in an image with Qwen Remove Lighting.
- [Qwen Shirt Design](https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/shirt-design): Mock up a graphic on a worn shirt with Qwen Shirt Design.
- [Qwen Image Max](https://raywake.si/explore/qwen-image-max/text-to-image): Create an image from text with Qwen Image Max.
- [Qwen Image Max Uncensored](https://raywake.si/explore/qwen-image-max-uncensored): Generate or edit images with the Wiro Qwen Image Max Uncensored operation.
- [Realistic Vision](https://raywake.si/explore/realistic-vision): Generate a photograph-like scene with Realistic Vision.
- [Recraft Vectorize](https://raywake.si/explore/recraft/vectorize): Convert a PNG, JPG or WebP image into a scalable SVG file.
- [Recraft Upscale](https://raywake.si/explore/recraft/upscale/crisp): Enlarge an existing image with the Crisp upscaler.
- [Recraft V3](https://raywake.si/explore/recraft/v3/text-to-image): Create a designed image from a prompt with Recraft V3.
- [Recraft V4](https://raywake.si/explore/recraft/v4/text-to-image): Develop an illustration or product scene from a written brief with preferred colors.
- [Recraft V4 Pro](https://raywake.si/explore/recraft/v4/pro/text-to-image): Build a composition with Recraft V4 Pro from a prompt and preferred palette.
- [Recraft V4 Custom Style](https://raywake.si/explore/recraft/v4/style/text-to-image): Generate images in a reusable custom style with Recraft V4 Standard or Pro.
- [Recraft V4.1 Flash](https://raywake.si/explore/recraft/v4.1/flash/text-to-image): Generate a raster image from text with Recraft V4.1 Flash.
- [Recraft V4.1](https://raywake.si/explore/recraft/v4.1/text-to-image): Generate a composed image from text with Recraft V4.1.
- [Recraft V4.1 Pro](https://raywake.si/explore/recraft/v4.1/pro/text-to-image): Create an image from a detailed brief with Recraft V4.1 Pro.
- [Recraft V4.1 Vector](https://raywake.si/explore/recraft/v4.1/text-to-vector): Create SVG illustrations and icons from a text prompt with optional color controls.
- [Rembg Enhance (Remove Background Enhance)](https://raywake.si/explore/smoretalk-ai/rembg-enhance): Remove a source image’s background with Rembg Enhance.
- [Remove Background](https://raywake.si/explore/imageutils/rembg): Create a subject cutout with the Remove Background utility.
- [Runway Gen-4 Turbo](https://raywake.si/explore/runway-gen4-turbo): Animate one image with Runway Gen-4 Turbo using a motion prompt.
- [SDXL ControlNet Union](https://raywake.si/explore/sdxl-controlnet-union): Combine structural controls for an SDXL scene with ControlNet Union.
- [SadTalker](https://raywake.si/explore/sadtalker): Turn a portrait and an audio recording into a talking-face video.
- [SAM 3.1 Image](https://raywake.si/explore/sam-3-1/image): Select objects in an image with SAM 3.1 segmentation masks.
- [Sana](https://raywake.si/explore/sana): Create images from a scene description with Sana.
- [Sana Sprint](https://raywake.si/explore/sana/sprint): Generate text-led images with the Sana Sprint variant.
- [Sana v1.5 1.6B](https://raywake.si/explore/sana/v1.5/1.6b): Generate images from text with the Sana v1.5 1.6B variant.
- [Sana v1.5 4.8B](https://raywake.si/explore/sana/v1.5/4.8b): Create images from a written brief with Sana v1.5 4.8B.
- [Seedance 1 Pro](https://raywake.si/explore/bytedance/seedance/v1/pro/text-to-video): Generate prompt-led video scenes using Seedance 1 Pro.
- [Seedance 2 Fast](https://raywake.si/explore/bytedance/seedance-2.0/fast/text-to-video): Generate short videos with optional synchronized sound using Seedance 2 Fast.
- [Seedance 2 Mini](https://raywake.si/explore/bytedance/seedance-2.0/mini/text-to-video): Create short prompt-led videos with Seedance 2 Mini.
- [Seedance 2](https://raywake.si/explore/bytedance/seedance-2.0/text-to-video): Create short videos with optional native audio using Seedance 2.
- [Seedance 2 US](https://raywake.si/explore/bytedance/seedance-2.0/us/text-to-video): Create a short video from a scene description, a starting image or image references with Seedance 2 US.
- [Seedance 2.5](https://raywake.si/explore/seedance-2-5): Move from a written scene to a video. Set the subject, action and mood in your prompt.
- [Seedance 2.5 US](https://raywake.si/explore/bytedance/seedance-2.5/us/text-to-video): Create videos from text, a starting frame or image references with the US-hosted Seedance 2.5 operation.
- [Seedream 5 Flash](https://raywake.si/explore/bytedance/seedream/v5/flash/text-to-image): Generate images or revise references with Seedream 5 Flash.
- [Seedream 5 Lite](https://raywake.si/explore/bytedance/seedream/v5/lite/text-to-image): Create images from text or edit supplied references with Seedream 5 Lite.
- [Seedream 4.5 Uncensored](https://raywake.si/explore/seedream-v4-5-uncensored): Generate or edit images with Wiro’s Seedream 4.5 Uncensored endpoint.
- [Seedream 5 Lite Uncensored](https://raywake.si/explore/seedream-v5-lite-uncensored): Create or revise images using Seedream 5 Lite Uncensored on Wiro.
- [Seedream 5 Pro](https://raywake.si/explore/bytedance/seedream/v5/pro/text-to-image): ByteDance's Seedream 5.0 Pro is flagship text-to-image model, with deep-thinking prompt understanding, native text in 14 languages, and precise control over dense layouts and structured designs.
- [Segment Anything 2](https://raywake.si/explore/sam2/image): Select regions in a still image with Segment Anything 2’s image mode.
- [SAM 3 Image](https://raywake.si/explore/sam-3/image): Select an object or region in an image to produce segmentation masks.
- [SenseNova U1 Infographic](https://raywake.si/explore/sensenova-u1-infographic): Turn an infographic brief into an image with SenseNova U1.
- [Stable Audio 2.5](https://raywake.si/explore/stable-audio-25/text-to-audio): Generate music or sound effects from a written audio brief.
- [Stable Audio 3 Medium](https://raywake.si/explore/stable-audio-3/medium/text-to-audio): Create, transform, repair or extend music and sound with Stable Audio 3 Medium.
- [Stable Audio 3 Small Music Base](https://raywake.si/explore/stable-audio-3/small/music/base/text-to-audio): Create, transform, repair or extend music with Stable Audio 3 Small Music Base.
- [Stable Audio 3 Medium Base](https://raywake.si/explore/stable-audio-3/medium/base/text-to-audio): Create, transform, repair or extend music and sound with Stable Audio 3 Medium Base.
- [Stable Audio 3 Small Music](https://raywake.si/explore/stable-audio-3/small/music/text-to-audio): Create, transform, repair or extend music with Stable Audio 3 Small Music.
- [Stable Audio 3 Small Sfx Base](https://raywake.si/explore/stable-audio-3/small/sfx/base/text-to-audio): Create, transform, repair or extend sound effects with Stable Audio 3 Small Sfx Base.
- [Stable Audio 3 Small Sfx](https://raywake.si/explore/stable-audio-3/small/sfx/text-to-audio): Create, transform, repair or extend sound effects with Stable Audio 3 Small Sfx.
- [Stable Diffusion 3.5 Medium](https://raywake.si/explore/stable-diffusion-v35-medium): Generate text-led images with Stable Diffusion 3.5 Medium.
- [Stable Diffusion 3 Medium](https://raywake.si/explore/stable-diffusion-v3-medium): Create images from text using Stable Diffusion 3 Medium.
- [Stable Diffusion XL](https://raywake.si/explore/fast-sdxl): Stable Diffusion XL generates images from text and supports optional LoRA weights.
- [SDXL Lightning](https://raywake.si/explore/fast-lightning-sdxl): SDXL Lightning generates images with a compact sampling schedule.
- [Stable Diffusion with LoRAs](https://raywake.si/explore/lora): Generate images with a chosen Stable Diffusion base model and LoRA adapters.
- [Stable Video Diffusion](https://raywake.si/explore/fast-svd/text-to-video): Stable Video Diffusion creates a short video from a scene description.
- [Stable Video Diffusion LCM](https://raywake.si/explore/fast-svd-lcm/text-to-video): Stable Video Diffusion LCM creates short clips from text, with a separate image animation mode.
- [Storm Trend Effect](https://raywake.si/explore/storm-trend-effect): Replace the central dancer in a preset Storm clip with the person from your portrait.
- [Subway Freeze Video](https://raywake.si/explore/subway-freeze-video): Make a subway freeze-frame style clip from a person photo.
- [T2V Turbo - Video Crafter](https://raywake.si/explore/t2v-turbo): Generate short clips from text with T2V Turbo using VideoCrafter.
- [Topaz Image Upscale](https://raywake.si/explore/topaz-upscale-image): Upscale images up to 4× with specialized enhancement models and optional face recovery.
- [ESRGAN Image Upscaler](https://raywake.si/explore/esrgan): Enlarge a source image with a selected RealESRGAN variant.
- [Veo 3.1](https://raywake.si/explore/veo3.1): Veo 3.1 generates video from a written scene, with optional audio and separate image-based operations.
- [Veo 3.1](https://raywake.si/explore/veo3.1/extend-video): Extend Veo-Created Videos up to 30 seconds
- [Veo 3.1 Fast](https://raywake.si/explore/veo-3-1-fast): Describe a scene and see it move. Explore short video ideas with atmosphere, camera movement and sound.
- [Veo 3.1 Fast](https://raywake.si/explore/veo3.1/fast/extend-video): Extend Veo-Created Videos up to 30 seconds
- [Veo 3.1 Lite](https://raywake.si/explore/veo3.1/lite): Veo 3.1 Lite generates short videos from text or a starting image, with audio that can be switched on or off.
- [Bria Video Background Removal](https://raywake.si/explore/bria/video/background-removal): Replace a clip’s background with the original Bria video-removal operation.
- [Bria Green Screen Despill](https://raywake.si/explore/bria/video/background-removal/green-screen-despill): Clean up a green-screen clip with Bria’s despill operation.
- [Bria Video Erase](https://raywake.si/explore/bria/video/erase/prompt): Remove a described object from a short video using Bria Video Erase.
- [Vidu Q2 Pro](https://raywake.si/explore/vidu/q2/image-to-video/pro): Vidu Q2 Pro animates a starting image; an optional ending image can define a transition. A separate operation extends video.
- [Vidu Q2 Turbo](https://raywake.si/explore/vidu/q2/image-to-video/turbo): Vidu Q2 Turbo turns a starting image into a short video, with an optional ending image for a transition.
- [Vidu Q2 Image](https://raywake.si/explore/vidu/q2/text-to-image): Vidu Q2 Image generates a still image from text, with a separate reference-image operation.
- [Vidu Q2](https://raywake.si/explore/vidu/q2/text-to-video): Vidu Q2 creates short video scenes from text, with selectable duration, framing and movement amplitude.
- [Vidu Reference to Image](https://raywake.si/explore/vidu/reference-to-image): Vidu Reference to Image creates a still image using a prompt and supplied reference images to guide the subject’s appearance.
- [Vidu Q1](https://raywake.si/explore/vidu/q1/text-to-video): Vidu Q1 creates videos from text, with separate operations for a starting image, reference images or two endpoint frames.
- [WAN 2.2 5B Image](https://raywake.si/explore/wan/v2.2-5b/text-to-image): WAN 2.2 5B Image creates still images from a written scene brief.
- [Wan 2.2 5B Fast](https://raywake.si/explore/wan/v2.2-5b/text-to-video/fast-wan): Wan 2.2 5B Fast turns a text description into a video scene.
- [WAN 2.2 A14B Image](https://raywake.si/explore/wan/v2.2-a14b/text-to-image): WAN 2.2 A14B Image creates images from text and offers a separate image editing mode.
- [Wan 2.2 A14B Turbo](https://raywake.si/explore/wan/v2.2-a14b/text-to-video/turbo): Wan 2.2 A14B Turbo creates video from text, with a separate image animation mode.
- [WAN 2.7 Pro Image](https://raywake.si/explore/wan/v2.7/pro/text-to-image): WAN 2.7 Pro Image creates images from text and includes a separate editing mode.
- [WAN 2.7 Image](https://raywake.si/explore/wan/v2.7/text-to-image): WAN 2.7 Image generates images from text or edits references in a separate mode.
- [WAN 2.5 Image](https://raywake.si/explore/wan-25-preview/text-to-image): WAN 2.5 Image generates still images from text, with a separate operation for editing supplied images.
- [Wan 2.5](https://raywake.si/explore/wan-25-preview/text-to-video): Wan 2.5 produces short videos from text, with optional audio guidance and a separate starting-image operation.
- [Wan 2.6 Uncensored](https://raywake.si/explore/wan-2-6-uncensored): Wan 2.6 Uncensored on Wiro generates video from text with optional starting-image and audio inputs.
- [Wan 2.7 Image](https://raywake.si/explore/wan-2-7-image): Wan 2.7 Image generates or edits images from a prompt, optionally drawing on up to nine input images.
- [Wan 2.7 Uncensored](https://raywake.si/explore/wan-2-7-uncensored): Wan 2.7 Uncensored on Wiro animates one required starting image using a written motion prompt and optional audio.
- [Wan 3.0](https://raywake.si/explore/alibaba/wan-3.0/image-to-video): Animate a starting image with Wan 3.0.
- [Wan 3.0](https://raywake.si/explore/wan-3-0): Wan 3.0 generates video from text, with optional first and last frame images and generated audio.
- [Wan 3.0 Prime](https://raywake.si/explore/alibaba/wan-3.0-prime/image-to-video): Animate a starting image with Wan 3.0 Prime.
- [Wan 3.0 Prime](https://raywake.si/explore/wan-3-0-prime): Wan 3.0 Prime turns a shot description into video, optionally guided by starting and ending images.
- [Wan 3.0 Prime Pro Uncensored](https://raywake.si/explore/wan-3-0-prime-pro-uncensored): Wan 3.0 Prime Pro Uncensored on Wiro creates videos with optional endpoint images, audio and output tiers through 4k.
- [Wan 3.0 Prime Uncensored](https://raywake.si/explore/wan-3-0-prime-uncensored): Wan 3.0 Prime Uncensored on Wiro generates text-led or image-anchored video, with optional audio and ending-frame guidance.
- [Wan 3.0 Pro Uncensored](https://raywake.si/explore/wan-3-0-pro-uncensored): Wan 3.0 Pro Uncensored on Wiro generates a shot from text or optional frame anchors, with 1080p, 2k and 4k output choices.
- [Wan 3.0 Uncensored](https://raywake.si/explore/wan-3-0-uncensored): Wan 3.0 Uncensored on Wiro creates video from a scene description, optionally using first and last images to guide continuity.
- [Wan 2.7](https://raywake.si/explore/wan/v2.7/text-to-video): Wan 2.7 generates video from text and offers image animation or video continuation separately.
- [Wan VACE 14B](https://raywake.si/explore/wan-vace-14b): Wan VACE 14B guides video generation or editing with source video, reference images, frames and masks.
- [Wan VACE 14B Depth](https://raywake.si/explore/wan-vace-14b/depth): Wan VACE 14B generates video using depth guidance and a scene prompt.
- [Wan VACE 14B Inpainting](https://raywake.si/explore/wan-vace-14b/inpainting): Wan VACE 14B regenerates masked regions within a source video.
- [Wan VACE 14B Outpainting](https://raywake.si/explore/wan-vace-14b/outpainting): Wan VACE 14B generates additional content around a source video’s edges.
- [Wan VACE 14B Pose](https://raywake.si/explore/wan-vace-14b/pose): Wan VACE 14B generates video guided by a source pose sequence.
- [Wan VACE 14B Reframe](https://raywake.si/explore/wan-vace-14b/reframe): Wan VACE 14B adapts source video framing with optional generation guidance.
- [Wan v2.2 A14B Text-to-Image A14B with LoRAs](https://raywake.si/explore/wan/v2.2-a14b/text-to-image/lora): Generate still images with WAN 2.2 A14B and optional LoRA weights.
- [Wan v2.6 Image to Image](https://raywake.si/explore/wan/v2.6/image-to-image): Wan 2.6 Image to Image edits or combines reference images using a text instruction.
- [WAN 2.6 Image](https://raywake.si/explore/wan/v2.6/text-to-image): Wan 2.6 text-to-image model.
- [Wan 2.6](https://raywake.si/explore/wan/v2.6/text-to-video): Wan 2.6 creates video from text or animates an image in a separate mode.
- [Wan 2.1](https://raywake.si/explore/wan-t2v): Wan 2.1 generates video from text, with separate operations for image animation and first/last-frame transitions.
- [Workflow Utilities Blend Video](https://raywake.si/explore/workflow-utilities/blend-video): Combine two clips as overlapping layers with Blend Video.
- [Workflow Utilities Extract Nth Frame](https://raywake.si/explore/workflow-utilities/extract-nth-frame): Extract regularly spaced still frames from an existing video for review or another image workflow.
- [Workflow Utilities Reverse Video](https://raywake.si/explore/workflow-utilities/reverse-video): Reverse an existing clip’s playback with the video utility.
- [Workflow Utilities Scale Video](https://raywake.si/explore/workflow-utilities/scale-video): Resize an existing video for a different layout.
- [Z Image Base](https://raywake.si/explore/z-image/base): Z Image Base generates still images from text with sampling controls.
- [Z Image Pro Uncensored](https://raywake.si/explore/z-image-pro-uncensored): Z Image Pro Uncensored creates an image from text through Wiro’s partner endpoint.
- [Z-Image Turbo](https://raywake.si/explore/z-image/turbo): Z-Image Turbo generates images from text, with separate transformation and masked editing modes.
- [GPT Image 1](https://raywake.si/explore/gpt-image-1/text-to-image): Generate images or edit references with GPT Image 1.

## Blog (English)

- [Raywake Blog](https://raywake.si/blog): Creative workflows and API guides.
- [RSS](https://raywake.si/blog/rss.xml): Full published articles.
- [AI Video API Costs: Quotes, Retries and Credit Settlement](https://raywake.si/blog/ai-video-api-costs): Budget AI video generation with quotes, credit holds and measured settlement. Learn how to avoid duplicate submissions and calculate cost per approved clip.
- [How to Turn an Image into a Video with AI, Step by Step](https://raywake.si/blog/image-to-video-ai): Turn a still image into a short AI video: pick the right frame, write a motion prompt, check the cost first and review the result. Studio and API versions.
- [One API for Image, Video and Audio: fal.ai vs Replicate vs Raywake](https://raywake.si/blog/one-api-image-video-audio): Compare fal.ai, Replicate and Raywake workflows, with four real perfume images from Nano Banana Pro, FLUX.2 Pro, Seedream 4.5 and GPT Image 2.5 Sunburst.
- [The Sora API Has Shut Down: Where to Move Your Video Workloads](https://raywake.si/blog/sora-api-alternative): OpenAI's Sora API and Videos API shut down on September 24, 2026 with no replacement. Here's how to pick a new video model by workload and migrate your code.

## Getting started

- [Raywake API](https://raywake.si/docs): Build image, video and audio generation with the Raywake API. Learn authentication, model inputs, credit quotes, job polling and result storage.
- [Quickstart](https://raywake.si/docs/quickstart): Make your first Raywake API request: choose a model, quote the input, start a generation and poll for results with curl, Python or JavaScript.
- [Authentication](https://raywake.si/docs/authentication): Authenticate with Raywake API keys, choose scopes and expiry, revoke access and keep credentials safe in your server-side integration.

## Core concepts

- [Quotes and credits](https://raywake.si/docs/credits): Understand Raywake credit quotes, generation holds, usage-based settlement, failed-job releases and the balance shared by studio and API.
- [Generations](https://raywake.si/docs/generations): Start and track Raywake generations. Learn job statuses, safe polling, idempotent retries, stored outputs and signed download links.
- [Inputs and files](https://raywake.si/docs/inputs): Read Raywake model input schemas, validate parameters and provide image, video or audio files for your generation requests.
- [Errors and limits](https://raywake.si/docs/errors): Handle Raywake API errors, permission failures, insufficient credits and rate limits. Learn status codes and when to retry a request.

## API reference

- [Models](https://raywake.si/docs/reference/models): List available Raywake AI models and read their input schemas. Explore model IDs, categories and the parameters required to generate media.
- [Quotes](https://raywake.si/docs/reference/quotes): Use the Raywake quotes endpoint to price an exact model and input before generation. Read quote IDs, credit costs and expiry times.
- [Create a generation](https://raywake.si/docs/reference/generate): Create image, video and audio jobs with the Raywake generation endpoint using a valid quote, model input and an idempotency key.
- [Jobs](https://raywake.si/docs/reference/jobs): Retrieve Raywake generation status and output links, or filter your job history by model, date, status and source through the API.
- [Wallet](https://raywake.si/docs/reference/wallet): Read your Raywake wallet through the API, including available credits, credits reserved for running jobs and any outstanding debt.
- [Usage](https://raywake.si/docs/reference/usage): Retrieve Raywake credit usage and generation activity by period, including daily totals, model breakdowns and failure counts.
- [Storage](https://raywake.si/docs/reference/storage): List Raywake uploads and generated outputs through the storage API. Read media types, file sizes, download links and deletion dates.
- [Pricing](https://raywake.si/docs/reference/pricing): Read published Raywake model credit rates through the pricing API, then request an exact-input quote before starting a generation.

## Resources

- [Model APIs](https://raywake.si/docs/model-apis): Find model-specific Raywake API examples and input schemas in the studio, with curl, Python and JavaScript generation workflows.
- [llms.txt](https://raywake.si/docs/llms): Use Raywake llms.txt to give AI assistants current product context, API workflow guidance, credit rules and links to the full documentation.

## Policies

- [Privacy](https://raywake.si/privacy): Data handling, retention and rights.
- [Terms](https://raywake.si/terms): Studio and API terms.
- [Refunds](https://raywake.si/refunds): Credit purchases, failed generations and withdrawal rights.
- [Acceptable use](https://raywake.si/acceptable-use): Content and activity rules.
- [Withdrawal](https://raywake.si/withdraw): Credit-purchase withdrawal request instructions.

# Model usage guidance



These notes are the same guidance shown on public model pages. Suggested prompts are examples, not measured comparisons or promises of a particular output. Check the current model schema and quote for your input before generating.



## AMT Interpolation
Source: https://raywake.si/explore/amt-interpolation
Model ID: amt-interpolation
Provider documentation: https://fal.ai/models/fal-ai/amt-interpolation/api
Content updated: 2026-10-01

Generate intermediate frames within an existing video. Use interpolation when you need to adjust frame delivery; it does not generate an unrelated scene from a prompt.

Inputs and settings:
- Provide video_url and choose output_fps. recursive_interpolation_passes controls repeated interpolation; check the current studio schema and quote before increasing it.

Review the result:
Play the result at its intended frame rate. Check fast movement, object crossings and shot changes for ghosting or distorted shapes, and compare timing with the source.

## Age Modification
Source: https://raywake.si/explore/image-apps-v2/age-modify
Model ID: image-apps-v2/age-modify
Provider documentation: https://fal.ai/models/fal-ai/image-apps-v2/age-modify/api
Content updated: 2026-10-01

Explore a different apparent age in a portrait with Age Modification. Use a clear face and compare the edit with the original.

Inputs and settings:
- Send image_url and choose target_age within the form limits.
- Use preserve_identity when retaining likeness matters; choose aspect_ratio for the portrait layout.

Review the result:
Check facial proportions and skin texture. This is a generated appearance, not a prediction of actual aging.

## AnimateDiff
Source: https://raywake.si/explore/fast-animatediff/text-to-video
Model ID: fast-animatediff/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/fast-animatediff/text-to-video/api
Content updated: 2026-10-01

Try a single visual idea with one clear subject action and restrained camera movement.

Inputs and settings:
- Choose num_frames from 1–32 and fps from 1–16 together; frame count and playback rate determine clip length.
- Use motions only for supported motion choices. Video-to-video instead needs video_url and prompt, with strength controlling the transformation.

Review the result:
Watch for flicker, changing faces and abrupt motion between the first and last frames.

## AnimateDiff Turbo
Source: https://raywake.si/explore/fast-animatediff/turbo/text-to-video
Model ID: fast-animatediff/turbo/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/fast-animatediff/turbo/text-to-video/api
Content updated: 2026-10-01

Preview a character action or a simple atmospheric scene.

Inputs and settings:
- The provider recommends 4–12 inference steps for Turbo. Set num_frames and fps together for the intended playback length.
- Keep guidance modest initially and refine the prompt first. The separate video-to-video mode requires video_url and exposes strength.

Review the result:
Check motion continuity and whether the subject changes shape during playback.

## AuraFlow
Source: https://raywake.si/explore/aura-flow
Model ID: aura-flow
Provider documentation: https://fal.ai/models/fal-ai/aura-flow/api
Content updated: 2026-10-01

Generate a new image from a written scene with AuraFlow. Describe subject, surroundings and light in a focused brief.

Inputs and settings:
- Send prompt; expand_prompt controls whether the provider expands the brief.
- Start with num_images set to one. Save seed and guidance_scale when comparing changes.

Review the result:
Check whether added details match the brief, especially when prompt expansion is enabled.

## AuraSR
Source: https://raywake.si/explore/aura-sr
Model ID: aura-sr
Provider documentation: https://fal.ai/models/fal-ai/aura-sr/api
Content updated: 2026-10-01

Upscale a photograph with AuraSR and compare checkpoint variants on the same source. Keep an untouched original for evaluating changed detail.

Inputs and settings:
- Provide image_url and select checkpoint from the available studio options.
- overlapping_tiles uses overlapping regions to help reduce seams. Compare the actual result and turnaround rather than assuming this option is always preferable.

Review the result:
Inspect tile boundaries, edges and fine texture. A larger image does not establish that newly added detail is accurate.

## BAGEL
Source: https://raywake.si/explore/bagel
Model ID: bagel
Provider documentation: https://fal.ai/models/fal-ai/bagel/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## BiRefNet V2 Video Background Removal
Source: https://raywake.si/explore/birefnet/v2/video
Model ID: birefnet/v2/video
Provider documentation: https://fal.ai/models/fal-ai/birefnet/v2/video/api
Content updated: 2026-10-01

Remove a video’s background with BiRefNet V2 and inspect the foreground across frames. Moving hair and objects need review throughout the clip.

Inputs and settings:
- Send video_url and select model and operating_resolution. General Use (Dynamic) is required for 2304x2304.
- Choose video_output_type, video_quality and video_write_mode; use output_mask for a separate mask video. Confirm transparency in the actual export rather than assuming every codec retains it.

Review the result:
Play the whole result and mask when requested. Check edge flicker, lost details and compatibility with your editing software.

## Birefnet Background Removal
Source: https://raywake.si/explore/birefnet
Model ID: birefnet
Provider documentation: https://fal.ai/models/fal-ai/birefnet/api
Content updated: 2026-10-01

Separate a foreground subject from its background. Keep the original and, when needed, request the mask so you can inspect or refine the selection.

Inputs and settings:
- Provide image_url and select model and operating_resolution in the studio. These controls affect how the source is processed rather than describing a new scene.
- Use output_mask to request the selection mask. Choose output_format and review refine_foreground before submitting.

Review the result:
Check hair, transparent objects and low-contrast edges over different backgrounds. Foreground refinement can change edge detail; compare it with the original before compositing.

## Birefnet Background Removal V2
Source: https://raywake.si/explore/birefnet/v2
Model ID: birefnet/v2
Provider documentation: https://fal.ai/models/fal-ai/birefnet/v2/api
Content updated: 2026-10-01

Separate a subject from its background with BiRefNet V2. Choose a variant for general images, portraits or matting before refining the cutout.

Inputs and settings:
- Provide image_url and choose model and operating_resolution. The 2304x2304 resolution requires General Use (Dynamic).
- Use output_mask to include a mask; mask_only returns the mask instead of a cutout. Select output_format and review refine_foreground.

Review the result:
Check hair, glass and gaps against light and dark backgrounds. Compare the selected foreground with the untouched source.

## Birthday Ride
Source: https://raywake.si/explore/birthday-ride
Model ID: birthday-ride
Content updated: 2026-10-02

Choose a clear, front-facing portrait of one person. The photo guides their appearance in the prepared celebration scene.

Inputs and settings:
- One 720p video costs 350 Raywake credits. Upload one portrait through image_url, review the credit quote, then generate.
- The preset produces a portrait 720p clip of about 8 seconds and retains the template’s original sound. The scene and timing are fixed, with no birthday text or black-screen ending.

Review the result:
Watch the face, hair and hands through the bottle spray and final kiss. Details can change during fast motion.

## BitDance
Source: https://raywake.si/explore/bitdance
Model ID: bitdance
Provider documentation: https://fal.ai/models/fal-ai/bitdance/api
Content updated: 2026-10-01

Explore photographic compositions by naming the subject, surface textures, lighting and background.

Inputs and settings:
- Provide prompt; adjust guidance_scale within 1–15 to explore adherence without changing the whole brief.
- Use num_images for 1–4 results and output_format for JPEG or PNG. Keep seed and the brief when comparing variations.

Review the result:
Inspect fine texture, object proportions and any lettering beyond the small preview.

## Boogu Image
Source: https://raywake.si/explore/boogu-image
Model ID: boogu-image
Provider documentation: https://fal.ai/models/fal-ai/boogu-image/api
Content updated: 2026-10-01

Try a detailed photographic or illustrated scene with clear placement and lighting.

Inputs and settings:
- Provide prompt and optionally negative_prompt. Explore guidance_scale from 1–8 and num_inference_steps from 20–50, changing one at a time.
- Use num_images for 1–4 outputs and output_format for JPEG or PNG. Leave cfg_range_start/end unchanged initially and retain seed for comparisons.

Review the result:
Check spatial relationships, repeated objects and details excluded in negative_prompt.

## Bria Upscale
Source: https://raywake.si/explore/bria/increase-resolution
Model ID: bria/increase-resolution
Provider documentation: https://fal.ai/models/bria/increase-resolution/api
Content updated: 2026-10-01

Increase an image’s resolution with Bria while controlling color and transparency preservation. Choose the enlargement for the intended delivery size.

Inputs and settings:
- Send image_url and choose desired_increase of 2 or 4. preserve_color corrects color cast against the source.
- Use preserve_alpha when keeping transparency and select PNG with output_type. JPEG does not retain an alpha channel.

Review the result:
Inspect sharp edges and color shifts against the source. Check transparent edges on both light and dark backgrounds.

## Bria Product Dimensions
Source: https://raywake.si/explore/bria/product-dimensions
Model ID: bria/product-dimensions
Provider documentation: https://fal.ai/models/bria/product-dimensions/api
Content updated: 2026-10-03

Add supplied physical dimension callouts to a product photo; the model does not measure the product for you.

Inputs and settings:
- Upload image and add at least one dimensions entry with a positive value and unit. Height, upper width and lower width are supported.
- Choose PNG, JPEG or dual (composite plus transparent overlay). Dual reserves a bound for two image outputs. Output size is 256–2200 px.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Background Remover
Source: https://raywake.si/explore/bria/background/remove
Model ID: bria/background/remove
Provider documentation: https://fal.ai/models/fal-ai/bria/background/remove/api
Content updated: 2026-10-01

Remove the background from an existing product or portrait photo. This operation takes an image rather than a scene prompt; keep the source available for comparison.

Inputs and settings:
- Provide the original through image_url. Review the exported image and its transparency before placing it into another design.

Review the result:
Inspect hair, glass, fine edges and gaps between objects. Place the cutout over both light and dark backgrounds to find remnants or missing foreground detail.

## Bria's VRMBG 3.0
Source: https://raywake.si/explore/bria/video/background-removal/v3
Model ID: bria/video/background-removal/v3
Provider documentation: https://fal.ai/models/bria/video/background-removal/v3/api
Content updated: 2026-10-01

Isolate a moving subject with Bria VRMBG 3.0. Decide whether the output needs a colored background or transparency before choosing the export format.

Inputs and settings:
- Provide video_url, choose background_color and output_container_and_codec, and set preserve_audio.
- auto_zoom crops to the smallest rectangle containing the subject throughout the clip. Check framing before using this crop in a fixed-size layout.

Review the result:
Check foreground continuity, fine edges and retained audio. Test transparency in your destination editor; an incompatible codec can produce a black background.

## Bytedance Omnihuman V1.5
Source: https://raywake.si/explore/bytedance/omnihuman/v1.5
Model ID: bytedance/omnihuman/v1.5
Provider documentation: https://fal.ai/models/fal-ai/bytedance/omnihuman/v1.5/api
Content updated: 2026-10-01

Build a speaking or performing character clip from a clear figure and matching audio.

Inputs and settings:
- Provide image_url and audio_url. With multiple people, optional mask_url uses white regions to identify the speaker.
- Choose resolution: audio must be under 30 seconds at 1080p or under 60 seconds at 720p. turbo_mode trades some quality for faster inference.

Review the result:
Compare mouth timing, gestures and face identity; listen for cut-off speech at the end.

## Seedance 1 Pro Fast
Source: https://raywake.si/explore/bytedance/seedance/v1/pro/fast/text-to-video
Model ID: bytedance/seedance/v1/pro/fast/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/bytedance/seedance/v1/pro/fast/text-to-video/api
Content updated: 2026-10-01

Explore a single action or a sequence with clear camera instructions.

Inputs and settings:
- Text mode requires prompt; image mode requires prompt and image_url. Use camera_fixed for a locked camera.
- Choose 2–12 seconds and 480p/720p/1080p. num_frames overrides duration when supplied, so use one timing approach deliberately; retain seed for comparison.

Review the result:
Check motion between frames, subject proportions and continuity through any cuts.

## Seedance 1.5 Pro
Source: https://raywake.si/explore/bytedance/seedance/v1.5/pro/text-to-video
Model ID: bytedance/seedance/v1.5/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/bytedance/seedance/v1.5/pro/text-to-video/api
Content updated: 2026-10-01

Develop a dialogue scene or moving shot with a clear action and concise sound direction.

Inputs and settings:
- Text mode needs prompt; image mode also needs image_url and accepts optional end_image_url.
- Choose 4–12 seconds and 480p/720p/1080p. Use camera_fixed for a stationary camera and generate_audio for sound; keep seed when comparing changes.

Review the result:
Check mouth movements, scene continuity and unwanted camera drift.

## Seedream 4
Source: https://raywake.si/explore/bytedance/seedream/v4/text-to-image
Model ID: bytedance/seedream/v4/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/bytedance/seedream/v4/text-to-image/api
Content updated: 2026-10-01

Explore a product scene, illustration or composition, then use editing mode to revise supplied images.

Inputs and settings:
- Text mode requires prompt. Editing also requires image_urls; describe the changes and what should remain.
- Start with one result using num_images, which accepts 1–6. Keep seed for comparison and inspect enhance_prompt_mode in the current form before changing it.

Review the result:
Check lettering, object count and source details that should survive an edit.

## Seedream 4.5
Source: https://raywake.si/explore/seedream-4-5
Model ID: bytedance/seedream/v4.5/text-to-image
Content updated: 2026-10-01

Explore architecture and landscape compositions from a written scene. The coastal-house example combines a large architectural form, an open horizon and warm evening light.

Suggested starting prompt (not the exact prompt behind the preview):
A concrete house above a rocky coast at sunset. Wide architectural composition, clean lines, warm interior lights and a calm ocean horizon.

Inputs and settings:
- The featured endpoint creates images from a required text prompt. To preserve or change an existing image, select its image-editing mode instead.
- Choose image_size for the layout and begin with one image. num_images counts separate generations; max_images can allow multiple outputs per generation, so they are not interchangeable.
- Inspect the returned images before increasing the batch size. Request a quote for the exact settings rather than assuming the same cost for every size or output count.

Review the result:
Review structural lines, perspective and reflections. Treat an architectural image as a visual concept, not a buildable plan or a factual rendering of a real property.

### How much does Seedream 4.5 cost per image on Raywake?
Request a Raywake quote with the exact image_size, num_images and other inputs. The returned credits apply to that request until the quote expires. The 7-credit image in our October 1 comparison is a dated sample with recorded settings; it is not a current price for every Seedream request.
Get a current quote: https://raywake.si/docs/reference/quotes

## Cartoonify
Source: https://raywake.si/explore/cartoonify
Model ID: cartoonify
Provider documentation: https://fal.ai/models/fal-ai/cartoonify/api
Content updated: 2026-10-01

Turn an existing photo into a cartoon treatment with Cartoonify. Begin with a subject that is easy to distinguish from its background.

Inputs and settings:
- Send image_url; this route starts from a photo rather than a text scene.
- Adjust scale for the effect strength. Keep seed fixed when comparing settings.

Review the result:
Compare the silhouette, expression and important accessories with the source.

## Chatterbox
Source: https://raywake.si/explore/chatterbox/text-to-speech
Model ID: chatterbox/text-to-speech
Provider documentation: https://fal.ai/models/fal-ai/chatterbox/text-to-speech/api
Content updated: 2026-10-03

Turn your text into speech with Chatterbox and its voice controls.

Inputs and settings:
- Enter non-empty text. The credit hold covers the complete text, including multibyte characters; final billing uses recorded provider usage.
- An optional reference audio guides the voice. Review temperature, exaggeration and CFG; HD also offers 48 kHz output.

Review the result:
Listen for names, pronunciation, pauses, clipping and the intended tone before publishing.

## Chatterboxhd
Source: https://raywake.si/explore/resemble-ai/chatterboxhd/text-to-speech
Model ID: resemble-ai/chatterboxhd/text-to-speech
Provider documentation: https://fal.ai/models/resemble-ai/chatterboxhd/text-to-speech/api
Content updated: 2026-10-03

Turn your text into speech with Chatterboxhd and its voice controls.

Inputs and settings:
- Enter non-empty text. The credit hold covers the complete text, including multibyte characters; final billing uses recorded provider usage.
- An optional reference audio guides the voice. Review temperature, exaggeration and CFG; HD also offers 48 kHz output.

Review the result:
Listen for names, pronunciation, pauses, clipping and the intended tone before publishing.

## Chrono Edit
Source: https://raywake.si/explore/chrono-edit
Model ID: chrono-edit
Provider documentation: https://fal.ai/models/fal-ai/chrono-edit/api
Content updated: 2026-10-01

Apply a written change to a source image with Chrono Edit. State the edit and the details that should stay intact.

Inputs and settings:
- Send image_url and prompt. enable_prompt_expansion controls expansion of the instruction.
- Choose resolution as 480p or 720p and select output_format. Quote again when changing resolution or turbo_mode.

Review the result:
Compare unchanged regions and subject identity with the source; inspect edges around the edited area.

## Chrono Edit LoRA
Source: https://raywake.si/explore/chrono-edit-lora
Model ID: chrono-edit-lora
Provider documentation: https://fal.ai/models/fal-ai/chrono-edit-lora/api
Content updated: 2026-10-01

Use the Chrono Edit LoRA variant for a source-image edit. Describe the intended change without assuming custom weights are available in the form.

Inputs and settings:
- Send image_url and prompt. The current Raywake form has no field for additional LoRA files.
- Choose resolution and output_format. Keep turbo_mode and seed in the record when comparing with the other Chrono route.

Review the result:
Check the actual edit against the source; do not attribute a result to a custom LoRA you did not supply.

## Chrono Edit Lora Gallery
Source: https://raywake.si/explore/chrono-edit-lora-gallery/paintbrush
Model ID: chrono-edit-lora-gallery/paintbrush
Provider documentation: https://fal.ai/models/fal-ai/chrono-edit-lora-gallery/paintbrush/api
Content updated: 2026-10-01

Sketch the shape or placement of a desired change and explain how that region should look.

Inputs and settings:
- Provide image_url and prompt. Optional mask_url marks sketch regions in black; do not reuse conventions from another masking model.
- Choose 480p or 720p and JPEG/PNG/WebP. Adjust lora_scale and guidance_scale separately; preserve seed while comparing the same sketch.

Review the result:
Compare edited boundaries, nearby textures and unmarked areas with the original.

## CogView
Source: https://raywake.si/explore/cogview4
Model ID: cogview4
Provider documentation: https://fal.ai/models/fal-ai/cogview4/api
Content updated: 2026-10-01

Specify subjects, their arrangement, lighting and materials for a scene with several visual requirements.

Inputs and settings:
- Provide prompt; use negative_prompt for specific exclusions. Adjust guidance_scale within 0–20 while holding the brief constant.
- Use num_inference_steps from 1–50 and num_images from 1–4. Choose JPEG or PNG with output_format and keep seed for comparisons.

Review the result:
Verify object count, spatial relationships and whether unwanted details remain.

## SDXL ControlNet Canny
Source: https://raywake.si/explore/fast-sdxl-controlnet-canny
Model ID: fast-sdxl-controlnet-canny
Provider documentation: https://fal.ai/models/fal-ai/fast-sdxl-controlnet-canny/api
Content updated: 2026-10-01

Guide an SDXL scene with edge structure using SDXL ControlNet Canny. Keep structural guidance separate from the appearance described in text.

Inputs and settings:
- Use control_image_url and prompt in the default mode; image-to-image and inpainting have their own input tables.
- controlnet_conditioning_scale sets structural influence. Keep seed fixed while adjusting it separately from guidance_scale.

Review the result:
Compare important contours and inspect areas where rigid guidance distorted the requested scene.

## Cosmos 3 Super Image
Source: https://raywake.si/explore/nvidia/cosmos-3-super/text-to-image
Model ID: nvidia/cosmos-3-super/text-to-image
Provider documentation: https://fal.ai/models/nvidia/cosmos-3-super/text-to-image/api
Content updated: 2026-10-01

Describe spatial relationships and materials for a scene concept.

Inputs and settings:
- This operation requires prompt and returns images. The wider Cosmos family supports other media, but this Raywake endpoint takes no source video or image.
- negative_prompt names unwanted features; guidance_scale controls prompt guidance. Set num_images and JPEG or PNG output_format, and record seed for comparisons.

Review the result:
Inspect object contact, reflections and any requested lettering.

## Cosmos Predict 2.5
Source: https://raywake.si/explore/cosmos-predict-2.5/text-to-video
Model ID: cosmos-predict-2.5/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/cosmos-predict-2.5/text-to-video/api
Content updated: 2026-10-01

Explore movement and object interactions from a written scene or supplied visual source.

Inputs and settings:
- Text mode requires prompt. Image mode adds image_url; video mode adds video_url. Match your source to the operation.
- Adjust guidance_scale from 1–20, optionally add negative_prompt and retain seed. Review video_output_type; Raywake exposes no duration or frame-count controls.

Review the result:
Check spatial continuity and contact between objects across frames; simulated motion is not a validated physics result.

## Cosmos Predict 2.5 Distilled
Source: https://raywake.si/explore/cosmos-predict-2.5/distilled/text-to-video
Model ID: cosmos-predict-2.5/distilled/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/cosmos-predict-2.5/distilled/text-to-video/api
Content updated: 2026-10-01

Describe one physical scene with a clear action to explore motion and spatial continuity.

Inputs and settings:
- Provide prompt and optionally negative_prompt to steer away from unwanted content. Retain seed when comparing the scene.
- Review video_output_type for export. Raywake exposes no duration, frame-count, sampling-step or guidance controls for this distilled operation.

Review the result:
Watch for impossible interactions, disappearing objects and flicker; generated footage is not physical evidence.

## Creative Upscaler
Source: https://raywake.si/explore/creative-upscaler
Model ID: creative-upscaler
Provider documentation: https://fal.ai/models/fal-ai/creative-upscaler/api
Content updated: 2026-10-01

Enlarge and reinterpret image detail with Creative Upscaler. Use it when some visual changes are acceptable.

Inputs and settings:
- Send image_url and select scale. Add prompt to describe the desired detail.
- Tune creativity and shape_preservation separately; compare against the original at full size.

Review the result:
Inspect faces, lettering and repeated textures for invented details.

## DeepSeek Janus-Pro
Source: https://raywake.si/explore/janus
Model ID: janus
Provider documentation: https://fal.ai/models/fal-ai/janus/api
Content updated: 2026-10-01

Create an image from a text brief with DeepSeek Janus-Pro. Specify composition before exploring variations.

Inputs and settings:
- Send prompt and choose output_format for delivery.
- temperature changes randomness; cfg_weight controls prompt guidance. Start with one image and quote before increasing num_images.

Review the result:
Check object relationships and small details against the brief.

## DreamOmni2
Source: https://raywake.si/explore/dreamomni2/edit
Model ID: dreamomni2/edit
Provider documentation: https://fal.ai/models/fal-ai/dreamomni2/edit/api
Content updated: 2026-10-01

Edit supplied images with a written instruction using DreamOmni2. Explain the role of each reference when combining several images.

Inputs and settings:
- Supply image_urls and prompt; both are required for this editing route.
- State what should change and what should remain. Keep reference order consistent with the instruction.

Review the result:
Check that the intended reference supplied the style or content and that preserved areas stayed recognizable.

## Dreamshaper
Source: https://raywake.si/explore/dreamshaper
Model ID: dreamshaper
Provider documentation: https://fal.ai/models/fal-ai/dreamshaper/api
Content updated: 2026-10-01

Explore illustrated or photographic concepts from text with Dreamshaper. Describe the intended medium as well as the subject.

Inputs and settings:
- Use prompt for the scene and negative_prompt for unwanted elements.
- Review the existing negative prompt before requesting illustration. Choose format and save seed for comparisons.

Review the result:
Check that exclusions do not conflict with the requested style; inspect anatomy and any lettering.

## Eleven v4
Source: https://raywake.si/explore/elevenlabs/tts/eleven-v4
Model ID: elevenlabs/tts/eleven-v4
Provider documentation: https://fal.ai/models/elevenlabs/tts/eleven-v4/api
Content updated: 2026-10-01

Read a voiceover script with Eleven v4, using audio tags to direct delivery. Keep spoken words and performance cues easy to distinguish.

Inputs and settings:
- Provide text and select voice. Use apply_text_normalization for numbers and dates, and language_code when you need a particular language.
- Adjust stability and similarity_boost separately. Select output_format for delivery; timestamps requests timing data, which you should check against the audio.

Review the result:
Listen for pronunciation, pauses and any cue read aloud. Check names and numbers against the script.

## Eleven v4 Turbo
Source: https://raywake.si/explore/elevenlabs/tts/eleven-v4-turbo
Model ID: elevenlabs/tts/eleven-v4-turbo
Provider documentation: https://fal.ai/models/elevenlabs/tts/eleven-v4-turbo/api
Content updated: 2026-10-01

Prepare an alternate narration with Eleven v4 Turbo. For a comparison with v4, reuse the same script, voice and delivery cues.

Inputs and settings:
- Send text, choose voice and record stability and similarity_boost. Keep normalization and language settings consistent across comparisons.
- Choose output_format for the destination. A seed requests best-effort repeatability; it does not guarantee identical audio.

Review the result:
Compare the complete audio for pronunciation and expression. Record actual turnaround and quotes rather than assuming an advantage from the Turbo name.

## ElevenLabs Audio Isolation
Source: https://raywake.si/explore/elevenlabs/audio-isolation
Model ID: elevenlabs/audio-isolation
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/audio-isolation/api
Content updated: 2026-10-01

Supply a recording to isolate speech from background sound. Raywake currently exposes audio_url for this operation; it does not expose the provider’s optional video input.

Inputs and settings:
- Provide the source recording through audio_url. Check the complete studio schema for accepted file constraints before submitting.

Review the result:
Listen to quiet words, breaths and pauses against the original. Removing background sound can also affect the voice; isolation does not transcribe or anonymize a speaker.

## ElevenLabs Dubbing
Source: https://raywake.si/explore/elevenlabs/dubbing
Model ID: elevenlabs/dubbing
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/dubbing/api
Content updated: 2026-10-01

Dub an existing audio recording into a selected language with ElevenLabs. Review the translated speech against the original message, not just its sound.

Inputs and settings:
- Provide audio_url and target_lang. This Raywake operation uses an audio source; a target language alone does not supply speech.
- Set source_lang when known; otherwise the provider detects it. num_speakers can specify the speaker count. Quote the source and settings before submitting.

Review the result:
Check names, numbers, meaning and speaker changes. For a video project, verify timing again after placing the dubbed audio in the edit.

## ElevenLabs TTS Multilingual v2
Source: https://raywake.si/explore/elevenlabs/tts/multilingual-v2
Model ID: elevenlabs/tts/multilingual-v2
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/tts/multilingual-v2/api
Content updated: 2026-10-01

Read narration in a supported language with ElevenLabs Multilingual v2. Write the final spoken script before adjusting delivery.

Inputs and settings:
- Provide text and choose voice and language_code. apply_text_normalization controls how numbers and dates are read.
- Adjust speed from 0.7 to 1.2, then compare stability, similarity_boost and style separately. Save the script and settings alongside the selected audio.

Review the result:
Listen for accents, proper names and numbers in the requested language. Check the whole recording before using timestamps for subtitles.

## ElevenLabs TTS Turbo v2.5
Source: https://raywake.si/explore/elevenlabs/tts/turbo-v2.5
Model ID: elevenlabs/tts/turbo-v2.5
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/tts/turbo-v2.5/api
Content updated: 2026-10-01

Adjust a spoken script’s pace with ElevenLabs Turbo v2.5. Separate a change in reading speed from a change in vocal expression.

Inputs and settings:
- Provide text and voice. speed accepts 0.7–1.2; choose a moderate starting value before testing the extremes.
- Record stability, similarity_boost and style. Use normalization for numbers or dates, and inspect requested timestamps before using them for captions.

Review the result:
Listen at normal playback speed. Check rushed words, unnatural pauses and the timing needed for your video.

## ElevenLabs Voice Changer
Source: https://raywake.si/explore/elevenlabs/voice-changer
Model ID: elevenlabs/voice-changer
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/voice-changer/api
Content updated: 2026-10-01

Transform a recorded performance with ElevenLabs Voice Changer. Start from a clear source recording and select the voice for the intended delivery.

Inputs and settings:
- Send audio_url and choose voice; this operation starts with recorded audio rather than a written script.
- remove_background_noise applies audio-isolation processing. Select output_format and record seed; judge noise removal from the actual result.

Review the result:
Compare words, timing and expression against the original. Check for altered pronunciation and unwanted noise throughout.

## ElevenLabs Music
Source: https://raywake.si/explore/elevenlabs-music
Model ID: elevenlabs/music
Content updated: 2026-10-01

Describe a musical direction for a scene or idea. Include the mood and instrumentation you want, then listen to the complete result before choosing where to use it.

Suggested starting prompt (not the exact prompt behind the preview):
A relaxed instrumental track with soft piano, warm bass and a light rhythm. Keep the mood reflective and leave room for spoken narration.

Inputs and settings:
- Choose either a descriptive prompt or a composition_plan for a structured arrangement. The prompt route uses music_length_ms for length in milliseconds; 30 seconds is 30,000 ms.
- With a prompt, use force_instrumental when you want music without vocals. For a composition plan, define sections and review respect_sections_durations rather than applying prompt-only controls.
- Select output_format from the current schema and quote the complete request. Listen to the beginning, transitions and ending; a musical prompt is not a text-to-speech script.

Review the result:
Listen through headphones and speakers. Check the beginning, transitions and ending; review the applicable terms for your intended use before publishing.

### Is ElevenLabs Music copyright-free?
A generation is not a blanket copyright-free license. ElevenLabs Music has applicable usage terms and restrictions on inputs that reference protected songs or artists. Check the terms and the rights to your inputs for the intended use; a generation price alone does not establish distribution rights.
ElevenLabs Music terms: https://elevenlabs.io/music-terms

## Elevenlabs Sound Effects V2
Source: https://raywake.si/explore/elevenlabs/sound-effects/v2
Model ID: elevenlabs/sound-effects/v2
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/sound-effects/v2/api
Content updated: 2026-10-01

Design a short sound effect from a description with ElevenLabs Sound Effects V2. Name the sound source, its action and the surrounding ambience.

Inputs and settings:
- Provide text and optionally duration_seconds from 0.5 to 22. Omitting duration lets the provider choose it from the description.
- prompt_influence controls adherence to the description. Choose loop and output_format, then quote the selected duration and settings.

Review the result:
Listen for unwanted voices, abrupt cuts and clipping. If requesting a loop, play several repetitions to check the join.

## Elevenlabs Tts Eleven V3
Source: https://raywake.si/explore/elevenlabs/tts/eleven-v3
Model ID: elevenlabs/tts/eleven-v3
Provider documentation: https://fal.ai/models/fal-ai/elevenlabs/tts/eleven-v3/api
Content updated: 2026-10-01

Create a spoken version of a script with Eleven v3. Keep pronunciation decisions in view when reviewing alternative deliveries.

Inputs and settings:
- Provide text, select voice and adjust stability. Use language_code for a supported language and normalization for number or date handling.
- Request timestamps when you need timing information. Check their alignment against the actual recording rather than assuming captions are ready to publish.

Review the result:
Listen for correct names, numbers and pauses. Compare the transcript with the audio before publishing or generating subtitles.

## Emu 3.5 Image
Source: https://raywake.si/explore/emu-3.5-image/text-to-image
Model ID: emu-3.5-image/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/emu-3.5-image/text-to-image/api
Content updated: 2026-10-01

Explore packaging concepts or illustrated scenes by specifying subject, light and composition.

Inputs and settings:
- Write a prompt and choose 480p or 720p resolution plus an aspect_ratio for the intended layout.
- Use output_format for JPEG, PNG or WebP. The separate editing mode requires image_url and a change instruction.

Review the result:
Inspect small lettering, object boundaries and whether the chosen crop keeps the subject visible.

## Ernie Image
Source: https://raywake.si/explore/ernie-image
Model ID: ernie-image
Provider documentation: https://fal.ai/models/fal-ai/ernie-image/api
Content updated: 2026-10-01

Draft scenes in English, Chinese or Japanese, with a negative prompt for unwanted details.

Inputs and settings:
- Start with one image. Adjust num_inference_steps within 1–100 and guidance_scale within 1–20 only after reviewing composition.
- Toggle enable_prompt_expansion to compare interpretation. Keep seed for comparisons and choose JPEG or PNG output.

Review the result:
Check whether expansion changed the intended subject, and review any generated lettering.

## Ernie Image Turbo
Source: https://raywake.si/explore/ernie-image/turbo
Model ID: ernie-image/turbo
Provider documentation: https://fal.ai/models/fal-ai/ernie-image/turbo/api
Content updated: 2026-10-01

Compare scene ideas or style directions before committing to a detailed composition.

Inputs and settings:
- The official Turbo starting point is eight inference steps. Raywake allows 1–20; begin with one image.
- Compare enable_prompt_expansion on and off. Use negative_prompt for exclusions and preserve seed when adjusting guidance_scale.

Review the result:
Inspect texture, facial features and whether important prompt details survived the shorter sampling.

## Bria Extract Object
Source: https://raywake.si/explore/bria/extract-object
Model ID: bria/extract-object
Provider documentation: https://fal.ai/models/bria/extract-object/api
Content updated: 2026-10-01

Select a particular object from a photograph with a written description. Use this extraction when a scene contains several possible foreground subjects.

Inputs and settings:
- Supply your own image_url and describe the target in prompt. autocrop tightens the output canvas around the extracted object.
- remove_background applies extra background-removal refinement. Compare its edges with the unrefined result, especially around lettering and graphics.

Review the result:
Inspect the cutout and mask against the original. Check thin details, holes, text and any neighboring object included by mistake.

## FLUX.2 [dev]
Source: https://raywake.si/explore/flux-2
Model ID: flux-2
Provider documentation: https://fal.ai/models/fal-ai/flux-2/api
Content updated: 2026-10-01

Build a new scene from text with FLUX.2 [dev]. Describe the subject, viewpoint and lighting before adjusting generation settings.

Inputs and settings:
- Use prompt in text-to-image mode. For a source photo, choose editing mode and check that mode’s input table.
- Adjust guidance_scale and num_inference_steps separately. enable_prompt_expansion lets the provider expand the brief.

Review the result:
Check perspective and object placement; compare added details with your original brief.

## FLUX.2 [flash]
Source: https://raywake.si/explore/flux-2/flash
Model ID: flux-2/flash
Provider documentation: https://fal.ai/models/fal-ai/flux-2/flash/api
Content updated: 2026-10-01

Explore text-based image variations with FLUX.2 [flash]. Keep one clear visual goal while changing the brief.

Inputs and settings:
- Use prompt for text generation; select the separate editing mode when supplying a photo.
- The current text form exposes guidance_scale and enable_prompt_expansion, but no inference-step setting. Save seed for comparisons.

Review the result:
Inspect material textures and lettering at full size before choosing a variation.

## FLUX 2 Lora Gallery Hdr Style
Source: https://raywake.si/explore/flux-2-lora-gallery/hdr-style
Model ID: flux-2-lora-gallery/hdr-style
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/hdr-style/api
Content updated: 2026-10-01

Create a dramatic tonal treatment from text with FLUX 2 HDR Style. Describe the lighting rather than expecting measured HDR capture.

Inputs and settings:
- Send prompt; the provider automatically adds the Hyp3rRe4list1c trigger.
- lora_scale controls the effect strength. Save seed while comparing shadow and highlight treatment.

Review the result:
Inspect halos, exaggerated local contrast and lost detail around bright edges.

## FLUX 2 Lora Gallery Realism
Source: https://raywake.si/explore/flux-2-lora-gallery/realism
Model ID: flux-2-lora-gallery/realism
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/realism/api
Content updated: 2026-10-01

Generate a photographic-style concept with FLUX 2 Realism. Describe ordinary lighting and surface details for the intended scene.

Inputs and settings:
- Send prompt; this style route generates a new image rather than editing a photo.
- Tune lora_scale independently of guidance_scale. Keep seed fixed when judging changes to the treatment.

Review the result:
Check reflections, skin and object proportions. Photographic appearance does not make the scene factual.

## FLUX 2 Turbo
Source: https://raywake.si/explore/flux-2/turbo
Model ID: flux-2/turbo
Provider documentation: https://fal.ai/models/fal-ai/flux-2/turbo/api
Content updated: 2026-10-01

Generate a fresh image from a scene brief with FLUX 2 Turbo. Write subject and framing explicitly rather than relying on the preview.

Inputs and settings:
- Send prompt. This listed route is text-to-image; it does not take a source-photo input.
- Choose output_format and start with num_images set to one. Keep enable_prompt_expansion consistent when comparing prompts.

Review the result:
Check the requested framing and whether prompt expansion introduced unwanted elements.

## FLUX.3 Image to Video
Source: https://raywake.si/explore/blackforestlabs/flux-3/image-to-video
Model ID: blackforestlabs/flux-3/image-to-video
Provider documentation: https://fal.ai/models/blackforestlabs/flux-3/image-to-video/api
Content updated: 2026-10-01

Use a finished composition as the opening frame and describe a continuous action.

Inputs and settings:
- Provide image_url and prompt. Select aspect_ratio, including auto, and 720p or 1080p resolution.
- Use generate_audio for sound. Raywake currently exposes no duration control for this operation; inspect the returned clip length before planning an edit.

Review the result:
Watch for changing facial features, distorted objects and sound that conflicts with movement.

## FLUX.1 Kontext [max]
Source: https://raywake.si/explore/flux-pro/kontext/max/text-to-image
Model ID: flux-pro/kontext/max/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/flux-pro/kontext/max/text-to-image/api
Content updated: 2026-10-01

Explore an alternative text-to-image composition with Kontext Max. For a comparison with Kontext Pro, reuse the same brief and frame rather than changing both the model and the scene.

Inputs and settings:
- Send prompt and choose aspect_ratio. This endpoint creates a new image; it does not accept a source photograph.
- Keep num_images, enhance_prompt, guidance_scale, seed and output_format consistent when comparing variants. Request a fresh quote for this endpoint.

Review the result:
Compare requested objects, spacing and text at the same displayed size. Keep the actual quotes and results; a model name is not evidence that it wins your specific task.

## FLUX.1 Kontext [pro]
Source: https://raywake.si/explore/flux-pro/kontext/text-to-image
Model ID: flux-pro/kontext/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/flux-pro/kontext/text-to-image/api
Content updated: 2026-10-01

Create a new image from a scene description with Kontext Pro. Describe the subject, composition and light; this text-to-image operation does not take a reference photo.

Inputs and settings:
- Provide prompt, choose aspect_ratio and begin with num_images set to 1. Select JPEG or PNG with output_format.
- Record enhance_prompt, guidance_scale and seed for comparisons. Change one control at a time and quote the exact image count.

Review the result:
Inspect lettering, object boundaries and reflections at full size. If a particular photograph must be preserved, choose an image-editing operation rather than describing it from memory.

## FLUX.1 Krea [dev]
Source: https://raywake.si/explore/flux-1/krea
Model ID: flux-1/krea
Provider documentation: https://fal.ai/models/fal-ai/flux-1/krea/api
Content updated: 2026-10-01

Explore a photographic scene from text with FLUX.1 Krea [dev]. Describe light, material and viewpoint in the brief.

Inputs and settings:
- Send prompt; this route is distinct from image-based Krea Redux.
- Compare guidance_scale and num_inference_steps one at a time. Preserve acceleration and seed in your comparison record.

Review the result:
Inspect skin or material texture and local lighting rather than treating a photographic style as proof of realism.

## FLUX.1 Krea [dev]
Source: https://raywake.si/explore/flux/krea
Model ID: flux/krea
Provider documentation: https://fal.ai/models/fal-ai/flux/krea/api
Content updated: 2026-10-01

Explore photographic treatments by describing believable light, textures and subject placement.

Inputs and settings:
- The provider’s starting point is 28 inference steps and guidance_scale 4.5. Use one image while refining the scene.
- Compare acceleration separately from prompt changes. Select JPEG or PNG output and keep seed when reviewing subtle lighting differences.

Review the result:
Look for over-smoothed skin, inconsistent reflections and missing scene details.

## FLUX.1 Krea [dev] Redux
Source: https://raywake.si/explore/flux-1/krea/redux
Model ID: flux-1/krea/redux
Provider documentation: https://fal.ai/models/fal-ai/flux-1/krea/redux/api
Content updated: 2026-10-01

Reinterpret a source image with FLUX.1 Krea [dev] Redux. Choose a reference with the visual qualities you want to explore.

Inputs and settings:
- Send image_url. This Redux form has no scene-prompt field.
- Save the source and seed together. Adjust guidance_scale separately from inference steps when comparing results.

Review the result:
Compare subject, palette and framing with the source; inspect unintended changes to faces or logos.

## FLUX.1 SRPO [dev]
Source: https://raywake.si/explore/flux-1/srpo
Model ID: flux-1/srpo
Provider documentation: https://fal.ai/models/fal-ai/flux-1/srpo/api
Content updated: 2026-10-01

Build a detailed visual brief for FLUX.1 SRPO [dev]. Define the subject’s pose and the scene’s lighting before generating.

Inputs and settings:
- Send prompt for a new image; the listed form has no source-image input.
- Record guidance_scale, seed and num_inference_steps. Choose JPEG or PNG through output_format.

Review the result:
Check pose, anatomy and lighting consistency; the model name is not evidence of quality on your particular request.

## FLUX.1 SRPO [dev]
Source: https://raywake.si/explore/flux/srpo
Model ID: flux/srpo
Provider documentation: https://fal.ai/models/fal-ai/flux/srpo/api
Content updated: 2026-10-01

Explore a portrait or editorial scene with explicit lighting, mood and material details.

Inputs and settings:
- Start with one image. The provider documents 28 inference steps and guidance_scale 4.5 as defaults.
- Compare one prompt or sampling change at a time with seed retained. Choose JPEG or PNG output for the intended use.

Review the result:
Judge composition alongside facial structure, fabric detail and any unwanted stylization.

## FLUX.1 [dev]
Source: https://raywake.si/explore/flux-1/dev
Model ID: flux-1/dev
Provider documentation: https://fal.ai/models/fal-ai/flux-1/dev/api
Content updated: 2026-10-01

Develop a text-to-image concept with FLUX.1 [dev]. Organize the brief around subject, environment and composition.

Inputs and settings:
- Use prompt for text generation; Redux is the separate image-reference route.
- The provider starts at 28 inference steps. Keep seed fixed and change guidance_scale or steps separately.

Review the result:
Check object count and relationships, especially when the brief contains several similar items.

## FLUX.1 [dev]
Source: https://raywake.si/explore/flux/dev
Model ID: flux/dev
Provider documentation: https://fal.ai/models/fal-ai/flux/dev/api
Content updated: 2026-10-01

Describe a scene’s subject, materials and lighting to explore a visual concept.

Inputs and settings:
- Text mode requires prompt. Begin with one image and adjust guidance_scale and num_inference_steps one at a time.
- Redux requires image_url instead of a text prompt. Choose JPEG or PNG output and retain seed and settings when comparing variants.

Review the result:
Inspect hands, object geometry and exact spelling of any requested text.

## FLUX.1 [dev] Redux
Source: https://raywake.si/explore/flux-1/dev/redux
Model ID: flux-1/dev/redux
Provider documentation: https://fal.ai/models/fal-ai/flux-1/dev/redux/api
Content updated: 2026-10-01

Generate a new interpretation of an image with FLUX.1 [dev] Redux. Use a clean reference whose subject is clearly visible.

Inputs and settings:
- Supply image_url instead of a text prompt. This route is not a targeted erase or mask edit.
- Use guidance_scale and num_inference_steps for comparisons; save the source URL and seed together.

Review the result:
Compare silhouette and composition with the source, then inspect new details at full size.

## FLUX.1 Differential Diffusion
Source: https://raywake.si/explore/flux-general/differential-diffusion
Model ID: flux-general/differential-diffusion
Provider documentation: https://fal.ai/models/fal-ai/flux-general/differential-diffusion/api
Content updated: 2026-10-01

Apply spatially guided changes when different parts of the source need different treatment.

Inputs and settings:
- Provide image_url, change_map_image_url and prompt. Start without extra adapters to see the map’s effect clearly.
- Tune strength within 0.01–1. Add compatible controlnets or loras separately, keeping seed fixed while comparing each change.

Review the result:
Inspect transitions around edited regions and compare areas meant to retain their original appearance.

## FLUX.1 RF-Inversion
Source: https://raywake.si/explore/flux-general/rf-inversion
Model ID: flux-general/rf-inversion
Provider documentation: https://fal.ai/models/fal-ai/flux-general/rf-inversion/api
Content updated: 2026-10-01

Try a targeted appearance change while assessing how closely the result follows the source.

Inputs and settings:
- Provide image_url and prompt. Keep the initial comparison free of extra loras or controlnets.
- controller_guidance_reverse values closer to one favor resemblance to the input. Adjust that setting independently and retain seed before adding further controls.

Review the result:
Compare facial identity, object placement and background detail against the input.

## FLUX.1 [schnell]
Source: https://raywake.si/explore/flux-1/schnell
Model ID: flux-1/schnell
Provider documentation: https://fal.ai/models/fal-ai/flux-1/schnell/api
Content updated: 2026-10-01

Sketch a scene from text with FLUX.1 [schnell]. Keep the description focused when exploring early visual directions.

Inputs and settings:
- Send prompt and start with the documented four-step default. The form allows 1–12 steps.
- Record acceleration and seed. Increase num_images only after quoting the complete request.

Review the result:
Check small subjects and text before using a draft in a final design; do not infer a fixed runtime from the name.

## FLUX.1 [schnell]
Source: https://raywake.si/explore/flux/schnell
Model ID: flux/schnell
Provider documentation: https://fal.ai/models/fal-ai/flux/schnell/api
Content updated: 2026-10-01

Draft several visual directions before selecting a composition to refine.

Inputs and settings:
- The official starting point is four steps. Raywake permits 1–12; compare step count with a fixed prompt and seed.
- Use num_images from 1–4 for variants. Redux is a separate image-led operation requiring image_url, with no prompt field.

Review the result:
Check whether fewer steps lost small objects, readable text or clear subject boundaries.

## FLUX.1 [schnell] Redux
Source: https://raywake.si/explore/flux-1/schnell/redux
Model ID: flux-1/schnell/redux
Provider documentation: https://fal.ai/models/fal-ai/flux-1/schnell/redux/api
Content updated: 2026-10-01

Explore image-led variations with FLUX.1 [schnell] Redux. Select a source with a clear subject and useful framing.

Inputs and settings:
- Send image_url; this form exposes neither prompt nor guidance_scale.
- Start at four inference steps and save seed. Choose output_format for the resulting image.

Review the result:
Compare subject boundaries and changed textures against the original; Redux need not reproduce it exactly.

## FLUX.2 [klein] 4B
Source: https://raywake.si/explore/flux-2/klein/4b
Model ID: flux-2/klein/4b
Provider documentation: https://fal.ai/models/fal-ai/flux-2/klein/4b/api
Content updated: 2026-10-01

Draft a text-led composition with FLUX.2 [klein] 4B. Make the main subject and its surroundings easy to distinguish.

Inputs and settings:
- Select text-to-image for prompt; editing uses its own input requirements.
- The text form allows 4–8 num_inference_steps. Start with the provider’s four-step default and save seed.

Review the result:
Check small objects and subject boundaries before expanding a draft into a finished asset.

## FLUX.2 [klein] 4B Base
Source: https://raywake.si/explore/flux-2/klein/4b/base
Model ID: flux-2/klein/4b/base
Provider documentation: https://fal.ai/models/fal-ai/flux-2/klein/4b/base/api
Content updated: 2026-10-01

Control a text-generated scene with FLUX.2 [klein] 4B Base. Separate the desired scene from details you want to exclude.

Inputs and settings:
- Use prompt and negative_prompt in text mode. Select editing mode for an existing image.
- Base exposes guidance_scale and 4–50 num_inference_steps; its provider default is 28 steps, unlike the four-step variant.

Review the result:
Check exclusions against the result; a negative prompt does not guarantee an element is absent.

## FLUX.2 [klein] 9B
Source: https://raywake.si/explore/flux-2/klein/9b
Model ID: flux-2/klein/9b
Provider documentation: https://fal.ai/models/fal-ai/flux-2/klein/9b/api
Content updated: 2026-10-01

Create a scene from a focused brief with FLUX.2 [klein] 9B. Describe spatial relationships when several subjects share the frame.

Inputs and settings:
- Use prompt for text mode; choose the separate editing mode to supply reference images.
- Start at the documented four-step default. This text form has a 4–8 step range and no guidance_scale control.

Review the result:
Inspect overlapping subjects and reflections. More model parameters alone do not establish better output for your brief.

## FLUX.2 [klein] 9B Base
Source: https://raywake.si/explore/flux-2/klein/9b/base
Model ID: flux-2/klein/9b/base
Provider documentation: https://fal.ai/models/fal-ai/flux-2/klein/9b/base/api
Content updated: 2026-10-01

Develop a scene with exclusions using FLUX.2 [klein] 9B Base. Keep the positive and negative descriptions consistent.

Inputs and settings:
- Use prompt and negative_prompt in text mode; inspect the editing schema separately.
- Base offers guidance_scale, acceleration and up to 50 inference steps. Record them together when comparing outputs.

Review the result:
Review fine structure and omissions at actual output size, not only the catalog thumbnail.

## FLUX1.1 [pro]
Source: https://raywake.si/explore/flux-pro/v1.1
Model ID: flux-pro/v1.1
Provider documentation: https://fal.ai/models/fal-ai/flux-pro/v1.1/api
Content updated: 2026-10-01

Create a new image from a written brief with FLUX1.1 [pro]. State the focal point and leave intentional space for your layout.

Inputs and settings:
- Send prompt; enhance_prompt controls provider enhancement of the instruction.
- Choose output_format as JPEG or PNG. Start with one image and obtain a new quote before increasing num_images.

Review the result:
Check composition, glass or metal reflections, and any generated packaging text.

## FLUX1.1 [pro] ultra
Source: https://raywake.si/explore/flux-pro/v1.1-ultra
Model ID: flux-pro/v1.1-ultra
Provider documentation: https://fal.ai/models/fal-ai/flux-pro/v1.1-ultra/api
Content updated: 2026-10-01

Explore a scene with optional image guidance using FLUX1.1 [pro] ultra. Distinguish visual inspiration from an exact image edit.

Inputs and settings:
- Send prompt. If using image_url, image_prompt_strength controls the source image’s influence.
- raw selects the documented less-processed treatment. Keep raw and enhance_prompt consistent when comparing variations.

Review the result:
Compare source resemblance and altered details; image guidance does not guarantee identity or exact composition.

## Fabric 1.0
Source: https://raywake.si/explore/veed/fabric-1.0
Model ID: veed/fabric-1.0
Provider documentation: https://fal.ai/models/veed/fabric-1.0/api
Content updated: 2026-10-01

Prepare a clear character image and the finished speech recording.

Inputs and settings:
- image_url and audio_url are both required. The recording supplies the performance; this operation has no text prompt or speech-generation field.
- resolution is required: choose 480p or 720p. Crop and clean the source materials beforehand, then review the complete video with sound enabled rather than only its preview frame.

Review the result:
Compare mouth movement with the audio and check facial stability across the clip.

## Face Retoucher
Source: https://raywake.si/explore/retoucher
Model ID: retoucher
Provider documentation: https://fal.ai/models/fal-ai/retoucher/api
Content updated: 2026-10-01

Retouch a face photo with Face Retoucher. Use a clear source portrait so you can judge the changes to skin and facial detail.

Inputs and settings:
- Send image_url; this route has no text instruction field.
- Save seed when comparing retouched versions of the same photo.

Review the result:
Inspect skin texture, eyes and facial shape; compare with the person’s original appearance.

## Feynobg Background Remover
Source: https://raywake.si/explore/feynobg
Model ID: feynobg
Provider documentation: https://fal.ai/models/fal-ai/feynobg/api
Content updated: 2026-10-01

Extract a foreground with Feynobg Background Remover. Choose a source with enough visible detail to inspect fine edges.

Inputs and settings:
- Send image_url; no text instruction is needed.
- Save seed when comparing cutouts. Keep the returned alpha channel in a format that supports transparency.

Review the result:
Inspect hair, holes and translucent edges against both light and dark backgrounds.

## Fibo
Source: https://raywake.si/explore/bria/fibo/generate
Model ID: bria/fibo/generate
Provider documentation: https://fal.ai/models/bria/fibo/generate/api
Content updated: 2026-10-03

Generate a composed image with Fibo, using a written or structured prompt.

Inputs and settings:
- Enter prompt or structured_prompt; an empty request cannot start. Set aspect_ratio and seed for a repeatable setup.
- Choose 1MP or 4MP.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Add Object by Text]
Source: https://raywake.si/explore/bria/fibo-edit/add_object_by_text
Model ID: bria/fibo-edit/add_object_by_text
Provider documentation: https://fal.ai/models/bria/fibo-edit/add_object_by_text/api
Content updated: 2026-10-03

Use Fibo Edit [Add Object by Text] to add an object at a described position.

Inputs and settings:
- Describe the new object and its position in instruction.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Colorize]
Source: https://raywake.si/explore/bria/fibo-edit/colorize
Model ID: bria/fibo-edit/colorize
Provider documentation: https://fal.ai/models/bria/fibo-edit/colorize/api
Content updated: 2026-10-03

Use Fibo Edit [Colorize] to change a photo’s color treatment.

Inputs and settings:
- Choose contemporary, vivid, black-and-white or sepia treatment.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Erase by Text]
Source: https://raywake.si/explore/bria/fibo-edit/erase_by_text
Model ID: bria/fibo-edit/erase_by_text
Provider documentation: https://fal.ai/models/bria/fibo-edit/erase_by_text/api
Content updated: 2026-10-03

Use Fibo Edit [Erase by Text] to remove a named object from a photo.

Inputs and settings:
- Name the object in object_name; no selection mask is needed.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Relight]
Source: https://raywake.si/explore/bria/fibo-edit/relight
Model ID: bria/fibo-edit/relight
Provider documentation: https://fal.ai/models/bria/fibo-edit/relight/api
Content updated: 2026-10-03

Use Fibo Edit [Relight] to change the direction and character of photo lighting.

Inputs and settings:
- Choose light_direction and light_type together.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Replace Object by Text]
Source: https://raywake.si/explore/bria/fibo-edit/replace_object_by_text
Model ID: bria/fibo-edit/replace_object_by_text
Provider documentation: https://fal.ai/models/bria/fibo-edit/replace_object_by_text/api
Content updated: 2026-10-03

Use Fibo Edit [Replace Object by Text] to replace an object using a written instruction.

Inputs and settings:
- State what to replace and its replacement in instruction.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Reseason]
Source: https://raywake.si/explore/bria/fibo-edit/reseason
Model ID: bria/fibo-edit/reseason
Provider documentation: https://fal.ai/models/bria/fibo-edit/reseason/api
Content updated: 2026-10-03

Use Fibo Edit [Reseason] to change the season in an existing scene.

Inputs and settings:
- Select spring, summer, autumn or winter.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Restore]
Source: https://raywake.si/explore/bria/fibo-edit/restore
Model ID: bria/fibo-edit/restore
Provider documentation: https://fal.ai/models/bria/fibo-edit/restore/api
Content updated: 2026-10-03

Use Fibo Edit [Restore] to restore a damaged or blurred photo.

Inputs and settings:
- Upload image_url; this operation needs no prompt.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Restyle]
Source: https://raywake.si/explore/bria/fibo-edit/restyle
Model ID: bria/fibo-edit/restyle
Provider documentation: https://fal.ai/models/bria/fibo-edit/restyle/api
Content updated: 2026-10-03

Use Fibo Edit [Restyle] to redraw a photo in a chosen artistic style.

Inputs and settings:
- Select one of the thirteen supplied styles.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Rewrite Text]
Source: https://raywake.si/explore/bria/fibo-edit/rewrite_text
Model ID: bria/fibo-edit/rewrite_text
Provider documentation: https://fal.ai/models/bria/fibo-edit/rewrite_text/api
Content updated: 2026-10-03

Use Fibo Edit [Rewrite Text] to replace lettering inside an image.

Inputs and settings:
- Enter the replacement wording in new_text.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Edit [Sketch to Image]
Source: https://raywake.si/explore/bria/fibo-edit/sketch_to_colored_image
Model ID: bria/fibo-edit/sketch_to_colored_image
Provider documentation: https://fal.ai/models/bria/fibo-edit/sketch_to_colored_image/api
Content updated: 2026-10-03

Use Fibo Edit [Sketch to Image] to turn a sketch into a colored image.

Inputs and settings:
- Upload the sketch; this operation needs no written prompt.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Gen 1.5 Text to Image
Source: https://raywake.si/explore/bria/fibo-gen-1.5/text-to-image
Model ID: bria/fibo-gen-1.5/text-to-image
Provider documentation: https://fal.ai/models/bria/fibo-gen-1.5/text-to-image/api
Content updated: 2026-10-03

Generate a composed image with Fibo Gen 1.5 Text to Image, using a written or structured prompt.

Inputs and settings:
- Enter prompt or structured_prompt; an empty request cannot start. Set aspect_ratio and seed for a repeatable setup.
- Choose 1MP or 4MP.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fibo Lite
Source: https://raywake.si/explore/bria/fibo-lite/generate
Model ID: bria/fibo-lite/generate
Provider documentation: https://fal.ai/models/bria/fibo-lite/generate/api
Content updated: 2026-10-03

Generate a composed image with Fibo Lite, using a written or structured prompt.

Inputs and settings:
- Enter prompt or structured_prompt; an empty request cannot start. Set aspect_ratio and seed for a repeatable setup.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Finegrain Eraser
Source: https://raywake.si/explore/finegrain-eraser
Model ID: finegrain-eraser
Provider documentation: https://fal.ai/models/fal-ai/finegrain-eraser/api
Content updated: 2026-10-01

Remove a described object from a photo with Finegrain Eraser. Name the object precisely when several similar subjects are present.

Inputs and settings:
- Send image_url and prompt describing what to erase.
- Choose mode from express, standard or premium; quote the selected mode before submitting.

Review the result:
Check the filled area, shadows and nearby objects for unwanted changes.

## Florence-2 Phrase Grounding
Source: https://raywake.si/explore/florence-2-large/caption-to-phrase-grounding
Model ID: florence-2-large/caption-to-phrase-grounding
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/caption-to-phrase-grounding/api
Content updated: 2026-10-01

Locate phrases from an image caption with Florence-2 phrase grounding. Supply a caption whose subjects can be checked in the source image.

Inputs and settings:
- Provide image_url and the caption in text_input.
- Inspect results.bboxes and labels together. Bounding boxes identify regions rather than precise object outlines or a finished image cutout.

Review the result:
Compare each returned region with the phrase it represents. Check missed subjects and incorrectly matched locations.

## Florence-2 Dense Region Captions
Source: https://raywake.si/explore/florence-2-large/dense-region-caption
Model ID: florence-2-large/dense-region-caption
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/dense-region-caption/api
Content updated: 2026-10-01

Describe individual image regions with Florence-2 dense region captions. Use these localized descriptions when reviewing a scene’s objects and relationships.

Inputs and settings:
- Send image_url without a prompt.
- Review results.bboxes and their labels together. Keep the source dimensions available when interpreting region coordinates.

Review the result:
Read each label alongside its region. Correct imagined attributes or relationships before turning the results into published captions or alt text.

## Florence-2 Object Detection
Source: https://raywake.si/explore/florence-2-large/object-detection
Model ID: florence-2-large/object-detection
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/object-detection/api
Content updated: 2026-10-01

Locate objects in a source image with Florence-2 object detection. Use the returned labels and bounding boxes as a starting point for image review.

Inputs and settings:
- Provide image_url; this operation does not need a text prompt.
- Inspect results.bboxes in the returned data. Detection boxes identify regions; they are not pixel-accurate cutout masks.

Review the result:
Compare each box with the original. Check missed objects, overlapping detections and incorrect labels before using the results downstream.

## Florence-2 OCR with Regions
Source: https://raywake.si/explore/florence-2-large/ocr-with-region
Model ID: florence-2-large/ocr-with-region
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/ocr-with-region/api
Content updated: 2026-10-01

Read text from an image together with its detected regions. Use the region output to relate recognized words to their location in the source.

Inputs and settings:
- Provide image_url. Inspect the returned text labels and region coordinates together; check the output schema before using them in an application.

Review the result:
Check names, numbers and punctuation against the original. Small lettering, rotation or a complex background can lead to omissions; OCR output needs review before reuse.

## Florence-2 Open-Vocabulary Detection
Source: https://raywake.si/explore/florence-2-large/open-vocabulary-detection
Model ID: florence-2-large/open-vocabulary-detection
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/open-vocabulary-detection/api
Content updated: 2026-10-01

Search an image for a named object with Florence-2 open-vocabulary detection. Choose a concrete object description that you can verify visually.

Inputs and settings:
- Send image_url and your object description in text_input.
- Inspect results.bboxes with their labels. If the target is ambiguous, revise the description and compare the new regions rather than accepting the first match.

Review the result:
Review every detection against the source. Check false matches and omissions; a model label does not establish that an object is present.

## Florence-2 Referring-Expression Segmentation
Source: https://raywake.si/explore/florence-2-large/referring-expression-segmentation
Model ID: florence-2-large/referring-expression-segmentation
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/referring-expression-segmentation/api
Content updated: 2026-10-01

Select an image region using a referring expression with Florence-2 segmentation. Describe which object you mean when several similar subjects are present.

Inputs and settings:
- Provide image_url and a locating expression in text_input, such as the object’s position and appearance.
- Read results.polygons rather than treating the output as detection boxes. This operation returns region coordinates, not an automatically finished transparent cutout.

Review the result:
Compare returned polygon boundaries with the intended object. Inspect thin parts and adjacent subjects before using the selection.

## Florence-2 Region Proposals
Source: https://raywake.si/explore/florence-2-large/region-proposal
Model ID: florence-2-large/region-proposal
Provider documentation: https://fal.ai/models/fal-ai/florence-2-large/region-proposal/api
Content updated: 2026-10-01

Find candidate regions in an image with Florence-2 region proposals. Use them to plan further inspection rather than assuming each proposal is a confirmed object.

Inputs and settings:
- Provide image_url; this operation does not require a text query.
- Inspect results.bboxes and their coordinates. Region proposals are candidate selections, not precise masks or verified object classifications.

Review the result:
Review overlapping and empty regions against the source. Choose only useful proposals for the next step of your workflow.

## Flow-Edit
Source: https://raywake.si/explore/flowedit
Model ID: flowedit
Provider documentation: https://fal.ai/models/fal-ai/flowedit/api
Content updated: 2026-10-01

Transform a photo using source and target descriptions with Flow-Edit. Describe the current scene accurately before writing the desired change.

Inputs and settings:
- Send image_url, source_prompt and target_prompt.
- Keep unchanged details consistent across both prompts. Adjust n_max for edit strength and compare with a saved seed.

Review the result:
Check whether the intended change happened without disturbing the composition you meant to keep.

## Flux 2 Flex
Source: https://raywake.si/explore/flux-2-flex
Model ID: flux-2-flex
Provider documentation: https://fal.ai/models/fal-ai/flux-2-flex/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Choose an image-size preset or enter custom dimensions within the provider limits. The quote accounts for the selected output size.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## FLUX 2 Ballpoint Pen Sketch
Source: https://raywake.si/explore/flux-2-lora-gallery/ballpoint-pen-sketch
Model ID: flux-2-lora-gallery/ballpoint-pen-sketch
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/ballpoint-pen-sketch/api
Content updated: 2026-10-01

Generate a ballpoint sketch from text with FLUX 2 Ballpoint Pen Sketch. Describe the subject and leave room for linework.

Inputs and settings:
- Include the documented b4llp01nt trigger in prompt.
- lora_scale sets style strength. Save seed and choose output_format before comparing variants.

Review the result:
Inspect line continuity, hatching and facial details at full size.

## FLUX 2 Digital Comic Art
Source: https://raywake.si/explore/flux-2-lora-gallery/digital-comic-art
Model ID: flux-2-lora-gallery/digital-comic-art
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/digital-comic-art/api
Content updated: 2026-10-01

Build a comic-style scene from text with FLUX 2 Digital Comic Art. Specify the action and framing for one clear composition.

Inputs and settings:
- Include the documented d1g1t4l trigger in prompt.
- Adjust lora_scale for the style influence; begin with num_images set to one.

Review the result:
Check pose readability and object boundaries. Inspect any speech or lettering separately.

## Flux 2 Lora Gallery
Source: https://raywake.si/explore/flux-2-lora-gallery/multiple-angles
Model ID: flux-2-lora-gallery/multiple-angles
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/multiple-angles/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Upload the source images here so their dimensions can be measured. Choose an output size or let it follow the first image. The quote accounts for all input images and the selected number of output images.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## FLUX 2 Satellite View Style
Source: https://raywake.si/explore/flux-2-lora-gallery/satellite-view-style
Model ID: flux-2-lora-gallery/satellite-view-style
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/satellite-view-style/api
Content updated: 2026-10-01

Create an imagined aerial scene with FLUX 2 Satellite View Style. Describe terrain, settlement patterns and overhead framing.

Inputs and settings:
- Send prompt; this route generates a scene rather than retrieving satellite imagery.
- Use lora_scale for style strength and save seed when comparing layouts.

Review the result:
Check road continuity and building patterns. Do not present the output as geographic evidence.

## FLUX 2 Sepia Vintage
Source: https://raywake.si/explore/flux-2-lora-gallery/sepia-vintage
Model ID: flux-2-lora-gallery/sepia-vintage
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/sepia-vintage/api
Content updated: 2026-10-01

Generate a sepia-toned photographic concept with FLUX 2 Sepia Vintage. Describe the scene and the period details you want.

Inputs and settings:
- Send prompt; this text route does not restore an uploaded historical photo.
- Adjust lora_scale for the vintage treatment and choose output_format for delivery.

Review the result:
Inspect clothing and props for inconsistent details. The generated scene is not a historical record.

## Flux 2 Lora Gallery
Source: https://raywake.si/explore/flux-2-lora-gallery/virtual-tryon
Model ID: flux-2-lora-gallery/virtual-tryon
Provider documentation: https://fal.ai/models/fal-ai/flux-2-lora-gallery/virtual-tryon/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Upload the source images here so their dimensions can be measured. Choose an output size or let it follow the first image. The quote accounts for all input images and the selected number of output images.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## Flux 2 Max
Source: https://raywake.si/explore/flux-2-max
Model ID: flux-2-max
Provider documentation: https://fal.ai/models/fal-ai/flux-2-max/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Choose an image-size preset or enter custom dimensions within the provider limits. The quote accounts for the selected output size.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## Flux 2 Max
Source: https://raywake.si/explore/flux-2-max/edit
Model ID: flux-2-max/edit
Provider documentation: https://fal.ai/models/fal-ai/flux-2-max/edit/api
Content updated: 2026-10-01

Combine reference details or revise a product scene while specifying what should stay unchanged.

Inputs and settings:
- Provide image_urls and prompt. Explain each image’s role and state the requested change directly.
- Raywake exposes JPEG or PNG output and seed. Refine the instruction when an edit drifts; this operation has no strength slider.

Review the result:
Compare reference identity, material texture and any labels at full size.

## FLUX.2 [pro]
Source: https://raywake.si/explore/flux-2-pro
Model ID: flux-2-pro
Content updated: 2026-10-01

Use a written brief to explore a product composition before arranging a physical shoot. The perfume example pairs a simple object with a textured surface and directional light.

Suggested starting prompt (not the exact prompt behind the preview):
A clear glass perfume bottle on pale travertine, soft side light, subtle reflections, a warm neutral background and generous space around the product.

Inputs and settings:
- This text-to-image endpoint requires a prompt. Describe the subject, lighting, background and framing; select an editing endpoint when an existing image must be part of the input.
- Use image_size for a preset composition or supported custom width and height. Set output_format to JPEG or PNG according to how you will use the image.
- Record the prompt, image_size and seed when comparing variations. Change one decision at a time, and inspect fine detail at the actual output size rather than judging only the preview.

Review the result:
Inspect the object edges, glass reflections and any lettering at full size. Generated packaging and logos may need correction before a commercial use.

### How is FLUX.2 Pro priced on Raywake?
Use the hosted FLUX.2 Pro operation with a current quote for your image size and other settings. Raywake generation uses prepaid credits. Local model weights, open-source variants and another provider’s token prices are separate products or billing arrangements.
Quote your image request: https://raywake.si/docs/reference/quotes

## FLUX.3 Video
Source: https://raywake.si/explore/blackforestlabs/flux-3/edit-video
Model ID: blackforestlabs/flux-3/edit-video
Provider documentation: https://fal.ai/models/blackforestlabs/flux-3/edit-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose 5–20 seconds, or auto where offered. Auto reserves for up to 20 seconds; final settlement uses provider-reported billing.
- Draft modes generate a 720p clip. Full generation offers 720p or 1080p. For keyframes, use unique frame_index values within duration × 24; up to ten keyframes are supported.
- Upload an MP4 source smaller than 50 MB and shorter than 15 seconds.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## FLUX.3 Image
Source: https://raywake.si/explore/blackforestlabs/flux-3/text-to-image
Model ID: blackforestlabs/flux-3/text-to-image
Provider documentation: https://fal.ai/models/blackforestlabs/flux-3/text-to-image/api
Content updated: 2026-10-03

Create an image from text or edit with up to ten ordered reference images using FLUX.3 Image.

Inputs and settings:
- Choose 512sq, 768sq, 1k, 2k or 4k and one of fifteen aspect ratios. JPEG and PNG are available.
- Edit mode takes one to ten uploaded images: each side must be at least 256 px and the area at most 4 MP. Refer to them as image 1, image 2, and so on.
- The reviewed launch tariff expires on October 8, 2026 at 15:00 UTC. The model pauses if that tariff expires or changes until its pricing is reviewed.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Fooocus Image-to-Image
Source: https://raywake.si/explore/fast-fooocus-sdxl/image-to-image
Model ID: fast-fooocus-sdxl/image-to-image
Provider documentation: https://fal.ai/models/fal-ai/fast-fooocus-sdxl/image-to-image/api
Content updated: 2026-10-01

Try a new visual treatment while comparing the result with the source image.

Inputs and settings:
- Provide image_url and prompt. Tune strength within 0.05–1; lower values generally retain more of the source.
- Compare enable_refiner separately. Use preserve_aspect_ratio and crop_output deliberately, and choose JPEG or PNG in format.

Review the result:
Check subject drift, altered edges and any crop that removes useful details.

## Fooocus
Source: https://raywake.si/explore/fooocus
Model ID: fooocus
Provider documentation: https://fal.ai/models/fal-ai/fooocus/api
Content updated: 2026-10-01

Try a visual treatment from a prompt, or start with an image when you need a local edit or a larger version. Choose the operation that matches the change you want.

Inputs and settings:
- Choose generation, inpainting or upscale/variation. Inpainting needs inpaint_image_url; upscale/variation needs uov_image_url. Use a mask for a specific region when the selected operation exposes it.
- Choose styles, performance and num_images within the current form. Record the seed when comparing changes, and choose output_format for delivery.
- The quote reserves a maximum of 300 compute seconds per run. The provider-recorded runtime determines the final charge; unused reserved credits return to your balance. This ceiling does not guarantee a completion time.

Review the result:
Inspect lettering, edges and fine detail. For inpainting, compare the masked boundary with the original; for upscaling, check that new details remain plausible.

## GLM Image
Source: https://raywake.si/explore/glm-image
Model ID: glm-image
Provider documentation: https://fal.ai/models/fal-ai/glm-image/api
Content updated: 2026-10-01

Describe each text line and its position when drafting menus or information cards.

Inputs and settings:
- Use prompt for the scene and num_images from 1–4. This route exposes no reference-image input.
- Compare num_inference_steps within 10–100 and guidance_scale within 1–10 separately. Disable enable_prompt_expansion when testing an exact brief.

Review the result:
Proofread lettering, numbers and spacing; inspect illustration details separately.

## GPT Image 1 Mini
Source: https://raywake.si/explore/gpt-image-1-mini
Model ID: gpt-image-1-mini
Provider documentation: https://fal.ai/models/fal-ai/gpt-image-1-mini/api
Content updated: 2026-10-01

Try layout and illustration ideas before choosing a final direction.

Inputs and settings:
- Choose image_size and quality; num_images accepts 1–4. Begin with one output to assess the brief.
- For transparency, use background with PNG or WebP output. Existing-image changes belong to edit mode, which requires image_urls.

Review the result:
Check small text, overlapping objects and cutout edges at the intended display size.

## GPT Image 2
Source: https://raywake.si/explore/openai/gpt-image-2
Model ID: openai/gpt-image-2
Provider documentation: https://fal.ai/models/openai/gpt-image-2/api
Content updated: 2026-10-01

Describe layout and write the exact wording required on signs or labels.

Inputs and settings:
- The default uses prompt. Editing additionally requires image_urls and accepts mask_url for regional edits; identify the parts that should remain unchanged.
- Select quality, num_images and output_format. For a transparent background, use background with PNG or WebP and check the exported image’s transparency.

Review the result:
Proofread text and inspect edges, repeated objects and edited regions.

## GPT Image 2.5 Flare
Source: https://raywake.si/explore/openai/gpt-image-2.5/flare/text-to-image
Model ID: openai/gpt-image-2.5/flare/text-to-image
Provider documentation: https://fal.ai/models/openai/gpt-image-2.5/flare/text-to-image/api
Content updated: 2026-10-01

Explore a layout, then review text and fine detail before increasing the output count.

Inputs and settings:
- Text-to-image requires prompt. Editing also requires image_urls and optionally mask_url; specify the intended changes and preserved areas.
- quality and num_images control the request. Choose output_format; output_compression applies only to JPEG or WebP. Use a transparency-capable format with transparent background.

Review the result:
Check typography, mask boundaries and transparency in the exported file.

## GPT Image 2.5
Source: https://raywake.si/explore/gpt-image-2-5
Model ID: openai/gpt-image-2.5/sunburst/edit
Content updated: 2026-10-01

Write a detailed visual brief for a food, product or editorial concept. Describe the composition and lighting together so that the result has a clear visual hierarchy.

Suggested starting prompt (not the exact prompt behind the preview):
A slice of cheesecake with fresh raspberries on a simple ceramic plate. Close editorial food photograph, soft studio light, realistic texture and an uncluttered background.

Inputs and settings:
- Sunburst has separate generation and editing operations. Text-to-image starts from prompt; editing requires image_urls and can also use a mask. Select the operation for your actual inputs.
- Choose quality and output_format, and begin with one image using num_images. Check the full schema for any additional controls, then quote the exact operation and settings.
- For a transparent background, use a supported output format with transparency. output_compression applies to JPEG or WebP; check fine lettering and edges after export.

Review the result:
Check fine texture, repeated objects and any requested text. For an edit, specify one change at a time so you can judge which instruction made the difference.

### How much does a GPT Image 2.5 Sunburst image cost on Raywake?
Quote the exact Sunburst operation, image dimensions, quality and count before generating. Raywake credits are shared by the studio and API. A ChatGPT subscription or a token price from a different model does not price this request.
Get a current quote: https://raywake.si/docs/reference/quotes

## GPT Image 1.5
Source: https://raywake.si/explore/gpt-image-1.5
Model ID: gpt-image-1.5
Provider documentation: https://fal.ai/models/fal-ai/gpt-image-1.5/api
Content updated: 2026-10-01

Specify subject placement, lighting and materials for a product or editorial concept.

Inputs and settings:
- Text mode offers three image_size values and low, medium or high quality. num_images ranges from 1–4.
- Edit mode requires image_urls and adds input_fidelity plus optional mask_image_url. Describe preserved details separately from the requested change.

Review the result:
Inspect requested details, text accuracy and transparency around fine edges.

## Gemini 2.5 Flash Image
Source: https://raywake.si/explore/gemini-25-flash-image
Model ID: gemini-25-flash-image
Provider documentation: https://fal.ai/models/fal-ai/gemini-25-flash-image/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Choose 1–4 images, an aspect ratio and PNG, JPEG or WebP. limit_generations stays enabled so the prompt cannot request additional unpriced images.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## Gemini 2.5 Flash Image
Source: https://raywake.si/explore/gemini-25-flash-image/edit
Model ID: gemini-25-flash-image/edit
Provider documentation: https://fal.ai/models/fal-ai/gemini-25-flash-image/edit/api
Content updated: 2026-10-01

Revise a background, combine references or change a subject’s appearance with explicit preservation instructions.

Inputs and settings:
- Provide image_urls and a prompt of at least three characters. Identify each reference’s role rather than giving conflicting directions.
- Choose aspect_ratio and JPEG, PNG or WebP output. Start with one image; this Raywake operation offers no resolution or strength control.

Review the result:
Compare source identity, boundaries and any text that the edit should retain.

## Gemini 3 Pro Image
Source: https://raywake.si/explore/gemini-3-pro-image-preview
Model ID: gemini-3-pro-image-preview
Provider documentation: https://fal.ai/models/fal-ai/gemini-3-pro-image-preview/api
Content updated: 2026-10-01

Draft an illustrated layout or detailed scene while specifying text and object placement.

Inputs and settings:
- Choose aspect_ratio and 1K, 2K or 4K resolution. Start with one image before increasing num_images.
- Use system_prompt for overall style instructions and prompt for the scene. Editing additionally requires image_urls; compare preserved details against the references.

Review the result:
Inspect spelling, object relationships and fine detail at the selected resolution.

## Gemini 3.1 Flash Image
Source: https://raywake.si/explore/gemini-3.1-flash-image-preview
Model ID: gemini-3.1-flash-image-preview
Provider documentation: https://fal.ai/models/fal-ai/gemini-3.1-flash-image-preview/api
Content updated: 2026-10-01

Explore banners or vertical designs with explicit subject placement and room for text.

Inputs and settings:
- Choose aspect_ratio, including 4:1, 1:4, 8:1 or 1:8, and resolution from 0.5K to 4K.
- Text generation starts with prompt. The separate edit mode accepts optional image_urls and other reference media; describe which details those references should inform.

Review the result:
Inspect edge cropping, small lettering and stretched-looking objects in extreme layouts.

## Gemini 3.1 Flash Tts
Source: https://raywake.si/explore/gemini-3.1-flash-tts
Model ID: gemini-3.1-flash-tts
Provider documentation: https://fal.ai/models/fal-ai/gemini-3.1-flash-tts/api
Content updated: 2026-10-01

Draft narration or spoken messages with separate instructions for tone and pace.

Inputs and settings:
- Put the spoken text in prompt and delivery direction in style_instructions. Choose voice and language_code for the intended reading.
- Select WAV, MP3 or Ogg Opus output. Compare temperature cautiously; this Raywake operation has no speakers configuration for assigning multiple voices.

Review the result:
Listen for pronunciation, unintended spoken instructions, pauses and clipped endings.

## Gemini 3.8 Flash Lite TTS
Source: https://raywake.si/explore/google/gemini-3.8-flash-lite-tts
Model ID: google/gemini-3.8-flash-lite-tts
Provider documentation: https://fal.ai/models/google/gemini-3.8-flash-lite-tts/api
Content updated: 2026-10-01

Use Flash Lite to read a short tutorial script or alternate between two configured speakers. When comparing it with Flash, keep the script, voices and delivery instructions the same.

Inputs and settings:
- Single-speaker speech uses prompt and voice; delivery instructions belong in style_instructions.
- Dialogue replaces prompt with exactly two speakers and ordered turns. Each turn needs a configured speaker_id and text; a turn can override the global style.

Review the result:
Compare pronunciation and pauses throughout. Measure turnaround time and quote; the Lite name alone does not establish a cost or quality advantage.

## Gemini 3.8 Flash TTS
Source: https://raywake.si/explore/google/gemini-3.8-flash-tts
Model ID: google/gemini-3.8-flash-tts
Provider documentation: https://fal.ai/models/google/gemini-3.8-flash-tts/api
Content updated: 2026-10-01

Prepare narration or a two-person conversation from a written script. Keep delivery directions separate from the words you want spoken.

Inputs and settings:
- For narration, send prompt and choose voice. Put pacing and tone in style_instructions.
- For dialogue, omit prompt. Configure exactly two speakers with distinct speaker_id values and voices, then supply ordered turns using those aliases and text.

Review the result:
Listen for pronunciation, speaker changes and unwanted spoken instructions. Check names and numbers against the script before attaching the audio to a video.

## Gemini Omni Flash
Source: https://raywake.si/explore/google/gemini-omni-flash/image-to-video
Model ID: google/gemini-omni-flash/image-to-video
Provider documentation: https://fal.ai/models/google/gemini-omni-flash/image-to-video/api
Content updated: 2026-10-01

Start from a clear image and describe subject movement, camera direction and ambient sound.

Inputs and settings:
- Send image_url and prompt. Select duration from 3–10 seconds and aspect_ratio as 16:9 or 9:16.
- Keep the prompt focused on movement already supported by the frame. This route has no resolution or end-frame field.

Review the result:
Watch for changing faces, bent objects and sound that does not match the action.

## Gemini Omni Flash 1.1
Source: https://raywake.si/explore/google/gemini-omni-flash/v1.1/text-to-video
Model ID: google/gemini-omni-flash/v1.1/text-to-video
Provider documentation: https://fal.ai/models/google/gemini-omni-flash/v1.1/text-to-video/api
Content updated: 2026-10-01

Write one scene with a clear action, camera movement and sound direction.

Inputs and settings:
- Text mode needs prompt; duration is 3–10 seconds. Choose 16:9 or 9:16 and a resolution from 360p to 4k.
- Image mode uses image_url and optional end_image_url. Reference mode accepts image_urls and reference_video_urls in its separate inputs.

Review the result:
Check motion continuity and whether audible events occur at the right moment.

## Ghiblify Images
Source: https://raywake.si/explore/ghiblify
Model ID: ghiblify
Provider documentation: https://fal.ai/models/fal-ai/ghiblify/api
Content updated: 2026-10-01

Explore an animated illustration treatment of a photo with Ghiblify Images. Choose a readable composition before applying the transformation.

Inputs and settings:
- Upload the source through image_url; this route does not accept a scene prompt.
- Save seed to compare variations while keeping the source photo unchanged.

Review the result:
Compare faces, clothing and background objects with the source rather than judging style alone.

## Google Veo 3
Source: https://raywake.si/explore/veo-3
Model ID: veo-3
Provider documentation: https://www.eachlabs.ai/google/veo3/veo-3
Content updated: 2026-10-01

Describe one scene with a clear action, camera move and sound direction.

Inputs and settings:
- prompt is required; this default operation starts from text. Choose aspect_ratio 16:9 or 9:16 and duration of 4, 6 or 8 seconds.
- Use generate_audio to choose sound. Raywake offers 720p and 1080p; the reviewed Eachlabs contract limits 1080p to eight seconds. Select settings before reviewing the quote.

Review the result:
Review motion continuity and listen for unintended dialogue or sound.

## Google Veo 3 Fast
Source: https://raywake.si/explore/veo-3-fast
Model ID: veo-3-fast
Provider documentation: https://www.eachlabs.ai/google/veo3/veo-3-fast
Content updated: 2026-10-01

Explore a scene idea with one action and a restrained camera move.

Inputs and settings:
- Describe the scene in prompt. Choose duration of 4, 6 or 8 seconds and aspect_ratio 16:9 or 9:16; this operation takes no source image.
- Set generate_audio deliberately. resolution supports 720p or 1080p, with 1080p restricted to eight seconds by the reviewed contract. The Fast label does not guarantee a completion time.

Review the result:
Check the whole clip for subject changes, flicker and audio that misses the brief.

## Google Veo 3 Fast Image to Video
Source: https://raywake.si/explore/veo-3-fast-image-to-video
Model ID: veo-3-fast-image-to-video
Provider documentation: https://www.eachlabs.ai/google/veo3/veo-3-fast-image-to-video
Content updated: 2026-10-01

Keep the image’s composition and describe the movement you want to add.

Inputs and settings:
- image_url and prompt are required. Specify subject motion and camera direction; choose aspect_ratio 16:9 or 9:16 and duration 4, 6 or 8 seconds.
- Select generate_audio for sound. resolution offers 720p and 1080p; the reviewed contract allows 1080p only at eight seconds. Review the clip with sound enabled when audio is requested.

Review the result:
Compare faces, product markings and opening composition with the source image.

## Google Veo 3 Image to Video
Source: https://raywake.si/explore/veo-3-image-to-video
Model ID: veo-3-image-to-video
Provider documentation: https://www.eachlabs.ai/google/veo3/veo-3-image-to-video
Content updated: 2026-10-01

Use it to animate a product photograph or storyboard frame with a clearly described camera move.

Inputs and settings:
- Supply image_url and prompt. Describe what moves, what stays still and any sound you want when generate_audio is enabled.
- Choose 16:9 or 9:16, 4, 6 or 8 seconds, and 720p or 1080p. Use 8 seconds for 1080p.

Review the result:
Check product markings, moving edges and whether sound matches the visible action.

## Google Virtual Try On
Source: https://raywake.si/explore/google/virtual-try-on
Model ID: google/virtual-try-on
Provider documentation: https://fal.ai/models/google/virtual-try-on/api
Content updated: 2026-10-01

Preview clothing on a person with Google Virtual Try On. Use a clear person photo and a separate image of the garment.

Inputs and settings:
- Send person_image_url and product_image_url in their respective fields.
- Start with num_images set to one; quote again before requesting more variations.

Review the result:
Inspect garment boundaries, prints and body proportions. A generated preview does not establish real fit or size.

## Grok Imagine Image
Source: https://raywake.si/explore/xai/grok-imagine-image
Model ID: xai/grok-imagine-image
Provider documentation: https://fal.ai/models/xai/grok-imagine-image/api
Content updated: 2026-10-01

Explore a visual direction with clear subject, composition and lighting instructions.

Inputs and settings:
- Text mode requires prompt. Choose aspect_ratio for the layout, resolution as 1k or 2k and num_images from 1–4.
- Select editing mode to provide image_urls. Choose output_format as jpeg, png or webp; use auto aspect ratio there when preserving the first image’s proportions.

Review the result:
Review faces, lettering and edge detail; compare edits against the original images.

## Grok Imagine Image Quality
Source: https://raywake.si/explore/xai/grok-imagine-image/quality/text-to-image
Model ID: xai/grok-imagine-image/quality/text-to-image
Provider documentation: https://fal.ai/models/xai/grok-imagine-image/quality/text-to-image/api
Content updated: 2026-10-01

Build an illustration or product concept from a complete composition brief.

Inputs and settings:
- Supply prompt; this operation accepts no reference images. Choose aspect_ratio, resolution as 1k or 2k and num_images from 1–4.
- Set output_format to jpeg, png or webp. The Quality endpoint has no separate quality or seed control in Raywake; compare results by changing the brief.

Review the result:
Inspect subject proportions, reflections and small text; compare variants at the same size.

## Grok Imagine Image 2.0
Source: https://raywake.si/explore/xai/grok-imagine-image/v2.0/text-to-image
Model ID: xai/grok-imagine-image/v2.0/text-to-image
Provider documentation: https://fal.ai/models/xai/grok-imagine-image/v2.0/text-to-image/api
Content updated: 2026-10-01

Test a composition before selecting output quality and resolution for delivery.

Inputs and settings:
- Provide prompt. Choose quality as low or medium separately from resolution, which supports 1k and 2k; num_images permits 1–4 outputs.
- Use aspect_ratio and output_format for delivery. Add image_urls only in editing mode, and describe both the change and the details to keep.

Review the result:
Compare fine detail and lettering at full size; in edits, inspect unintended changes.

## Grok Imagine Video
Source: https://raywake.si/explore/xai/grok-imagine-video/text-to-video
Model ID: xai/grok-imagine-video/text-to-video
Provider documentation: https://fal.ai/models/xai/grok-imagine-video/text-to-video/api
Content updated: 2026-10-01

Describe one action, a camera direction and the intended sound for a short scene.

Inputs and settings:
- Default mode requires prompt. Select duration from 1–15 seconds, resolution as 480p or 720p and aspect_ratio for delivery.
- Image-to-video requires image_url and prompt. Match the source composition to the intended shot and describe movement rather than replacing the whole scene.

Review the result:
Watch the whole clip with sound; check timing, subject continuity and distracting background changes.

## Grok Imagine Video 1.5
Source: https://raywake.si/explore/xai/grok-imagine-video/v1.5/text-to-video
Model ID: xai/grok-imagine-video/v1.5/text-to-video
Provider documentation: https://fal.ai/models/xai/grok-imagine-video/v1.5/text-to-video/api
Content updated: 2026-10-01

Develop a short shot with a defined action, camera move and sound direction.

Inputs and settings:
- Text mode requires prompt. Choose duration from 1–15 seconds and resolution from 480p, 720p or 1080p; select aspect_ratio for a new shot.
- Image mode requires image_url and prompt. Raywake exposes no aspect_ratio there, so prepare the source framing before upload and describe its movement.

Review the result:
Review sound timing and facial detail; watch for changes in the subject across frames.

## MiniMax H3 Max Styles
Source: https://raywake.si/explore/minimax/h3-max/styles/16bit-pixel
Model ID: minimax/h3-max/styles/16bit-pixel
Provider documentation: https://fal.ai/models/minimax/h3-max/styles/16bit-pixel/api
Content updated: 2026-10-01

The default applies pixel styling automatically. Separate hand-drawn, low-poly, retro-toon and VHS operations provide other looks.

Inputs and settings:
- prompt is required; duration ranges from 5–15 seconds. For text generation, choose aspect_ratio; an optional image_url supplies the opening frame.
- When image_url is supplied, its ratio determines the canvas. Start with pixel artwork for style consistency; damage_level belongs only to the VHS operation.

Review the result:
Check pixel-edge consistency during movement and listen for sound matching the scene.

## MiniMax H3 Max Recast
Source: https://raywake.si/explore/minimax/h3-max/recast
Model ID: minimax/h3-max/recast
Provider documentation: https://fal.ai/models/minimax/h3-max/recast/api
Content updated: 2026-10-02

Upload a source clip and one photo per new person. References replace the main people from left to right; add an optional prompt to direct who replaces whom.

Inputs and settings:
- Source video must be 5–30 seconds long, with no single shot longer than 15 seconds. Upload it so its duration can be priced before generation.
- Supply 1–4 reference_image_urls. Choose 768P or 1080P; prompt is optional up to 2000 characters and seed is optional. Reference photos carry no additional charge.

Review the result:
Compare identity, facial details, motion, camera movement, cuts and audio against the source throughout the video.

## MiniMax H3 Max Turbo
Source: https://raywake.si/explore/minimax/h3-max-turbo/text-to-video
Model ID: minimax/h3-max-turbo/text-to-video
Provider documentation: https://fal.ai/models/minimax/h3-max-turbo/text-to-video/api
Content updated: 2026-10-01

Describe action, camera and sound together. Use the image operation for an opening still or a first-to-last-frame transition.

Inputs and settings:
- prompt and prompt_expansion_mode are required. Choose duration within 5–15 seconds and resolution from 480P, 768P or 1080P.
- target_audio_url replaces the soundtrack: a longer clip is trimmed and a shorter one padded with silence. Keep seed fixed when comparing prompt changes.

Review the result:
Review motion, lettering and the soundtrack throughout; do not judge a video from its poster.

## Happy Horse
Source: https://raywake.si/explore/alibaba/happy-horse/text-to-video
Model ID: alibaba/happy-horse/text-to-video
Provider documentation: https://fal.ai/models/alibaba/happy-horse/text-to-video/api
Content updated: 2026-10-01

Explore an atmospheric scene or a brief sequence; describe actions, camera movement and sounds together.

Inputs and settings:
- Text mode requires prompt. Reference mode also requires image_urls; use that operation when supplied images should guide the scene.
- Choose duration from 3–15 seconds, resolution at 720p or 1080p, and aspect_ratio for the destination. Retain seed when comparing revisions.

Review the result:
Watch for drifting faces, awkward cuts and sounds that arrive before or after the action.

## Happy Horse 1.1
Source: https://raywake.si/explore/alibaba/happy-horse/v1.1/text-to-video
Model ID: alibaba/happy-horse/v1.1/text-to-video
Provider documentation: https://fal.ai/models/alibaba/happy-horse/v1.1/text-to-video/api
Content updated: 2026-10-01

Plan dialogue or ambient scenes with explicit actions and camera direction.

Inputs and settings:
- Text mode requires prompt. Image mode requires image_url and makes prompt optional; reference mode requires image_urls and prompt.
- Select 3–15 seconds and 720p or 1080p. Text and reference modes expose aspect_ratio; the single-image operation does not.

Review the result:
Review speech timing, character continuity and the first and last frames.

## Hidream O1 Image
Source: https://raywake.si/explore/hidream-o1-image
Model ID: hidream-o1-image
Provider documentation: https://fal.ai/models/fal-ai/hidream-o1-image/api
Content updated: 2026-10-01

Describe shape, materials and lighting for a scene or product concept.

Inputs and settings:
- Use prompt; this public route has no reference-image input. num_images ranges from 1–4.
- The full model defaults to 50 num_inference_steps and guidance_scale 5 upstream. Compare one change at a time and save seed.

Review the result:
Check object geometry and whether fine texture became noise.

## Hidream O1 Image
Source: https://raywake.si/explore/hidream-o1-image/dev
Model ID: hidream-o1-image/dev
Provider documentation: https://fal.ai/models/fal-ai/hidream-o1-image/dev/api
Content updated: 2026-10-01

Use a stable scene brief when comparing the distilled Dev route with the full model.

Inputs and settings:
- Send prompt and begin with num_images 1. Raywake exposes no reference-image field here.
- Dev is documented for 28 num_inference_steps and guidance_scale 0. The public step range ends at 28; keep seed for comparisons.

Review the result:
Inspect edges, anatomy and texture rather than comparing appearance alone.

## Hotel Lobby Studio Video
Source: https://raywake.si/explore/hotel-lobby-studio-video
Model ID: hotel-lobby-studio-video
Provider documentation: https://www.eachlabs.ai/eachlabs/eachlabs/hotel-room
Content updated: 2026-10-01

Choose clear portraits with visible faces for the left and optional right performer.

Inputs and settings:
- Supply user_photo; user_photo_2 adds the second person. This template has no free-text prompt.
- Select duration from 4–15 seconds and resolution as 480p, 720p or 1080p. Preview the choreography before choosing a longer output.

Review the result:
Watch the full clip for identity drift, face changes and mismatched mouth movement.

## Hunyuan Video (Video-to-Video)
Source: https://raywake.si/explore/hunyuan-video/video-to-video
Model ID: hunyuan-video/video-to-video
Provider documentation: https://fal.ai/models/fal-ai/hunyuan-video/video-to-video/api
Content updated: 2026-10-01

Keep the source motion as a starting point while describing a new visual treatment.

Inputs and settings:
- Send video_url and prompt. Adjust strength within 0.01–1; compare a modest change before pushing the transformation further.
- Choose resolution from 480p, 580p or 720p. pro_mode uses more processing upstream; num_frames controls frames, not a seconds field.

Review the result:
Compare motion timing, subject identity and background stability against the original.

## Hunyuan World
Source: https://raywake.si/explore/hunyuan_world
Model ID: hunyuan_world
Provider documentation: https://fal.ai/models/fal-ai/hunyuan_world/api
Content updated: 2026-10-01

Describe the surrounding environment so the source fits into a wider scene.

Inputs and settings:
- Send image_url and prompt describing the panorama.
- This public route exposes only those two fields. Adjust the source and scene description rather than assuming 3D export controls are available.

Review the result:
Inspect the horizon, repeated structures and the join between panorama edges.

## Ideogram Custom Models
Source: https://raywake.si/explore/ideogram/custom-models/generate
Model ID: ideogram/custom-models/generate
Provider documentation: https://fal.ai/models/fal-ai/ideogram/custom-models/generate/api
Content updated: 2026-10-01

Use a completed custom model when a learned subject or visual identity should guide new scenes.

Inputs and settings:
- Supply model_id and prompt. model_id must identify a model whose training has completed.
- Select aspect_ratio and begin with num_images 1. Compare rendering_speed and expand_prompt while preserving the same model and seed.

Review the result:
Check whether the learned subject survives changes in viewpoint and lighting.

## Ideogram V3 Layerize Text
Source: https://raywake.si/explore/ideogram/v3/layerize-text
Model ID: ideogram/v3/layerize-text
Content updated: 2026-10-01

Use a flat graphic when you need to separate its lettering from the underlying image. Layerize Text produces an image and structured text information for editing or recomposition; it is not the same task as generating a new design from a prompt.

Inputs and settings:
- Provide the graphic through image_url. Begin with clear, legible text and keep the original graphic for comparison.
- Review the returned image and structured text together. Check the current output schema before importing the result into your design or application.

Review the result:
Compare the extracted words, line breaks and placement with the source. Review the image underneath the removed text for lost detail before rebuilding the design.

## Ideogram
Source: https://raywake.si/explore/ideogram/v3/reframe
Model ID: ideogram/v3/reframe
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v3/reframe/api
Content updated: 2026-10-03

Reframe an existing image into a new canvas with Ideogram V3.

Inputs and settings:
- Upload image_url, choose image_size, TURBO/BALANCED/QUALITY and 1–8 outputs. Each output is included in the quote.
- Style references must be uploaded JPEG, PNG or WebP files totaling no more than 10 MB. Eight-character hexadecimal style_codes cannot be combined with style or reference images.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Ideogram Object Removal
Source: https://raywake.si/explore/ideogram/object-removal
Model ID: ideogram/object-removal
Provider documentation: https://fal.ai/models/fal-ai/ideogram/object-removal/api
Content updated: 2026-10-01

Remove a selected object from an existing image using a mask. Select the unwanted object carefully so nearby surfaces and subjects remain outside the removal region.

Inputs and settings:
- Provide image_url and mask_url; no written prompt is required. The black-and-white mask must match the source dimensions.
- For this object-removal operation, white pixels mark removal and black pixels mark preservation. Do not reuse another editing model’s mask convention without checking it.

Review the result:
Inspect the filled region, shadows and repeating textures against the original. Removing an object can leave an implausible surface or remnants around its edge.

## Ideogram Remove Background
Source: https://raywake.si/explore/ideogram/remove-background
Model ID: ideogram/remove-background
Provider documentation: https://fal.ai/models/fal-ai/ideogram/remove-background/api
Content updated: 2026-10-01

Prepare a foreground cutout from a product or portrait photo. Use the resulting transparent image when placing the subject onto a different background.

Inputs and settings:
- Send image_url with a JPEG, PNG or WebP source; the provider specifies a maximum file size of 10 MB. This operation does not need a prompt or selection mask.
- Keep an export with transparency. Converting the cutout to JPEG removes its alpha channel; inspect the final composite separately.

Review the result:
View the cutout against light and dark backgrounds. Check hair, glass, gaps and contact shadows for missing foreground or retained background.

## Ideogram V3
Source: https://raywake.si/explore/ideogram/v3
Model ID: ideogram/v3
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v3/api
Content updated: 2026-10-01

Use a clear text brief and optional image_urls for style references.

Inputs and settings:
- Choose rendering_speed separately; num_images accepts 1–8. Use style_codes without style or image style references.
- Edit requires image_url and mask_url; remix adds strength. Background replacement and transparent generation have their own inputs.

Review the result:
Proofread text and check that reference styling has not obscured the subject.

## Ideogram Upscale
Source: https://raywake.si/explore/ideogram/upscale
Model ID: ideogram/upscale
Provider documentation: https://fal.ai/models/fal-ai/ideogram/upscale/api
Content updated: 2026-10-01

Enlarge an existing image while reviewing the balance between source resemblance and added detail. Keep an untouched source for checking packaging, faces and fine text.

Inputs and settings:
- Provide image_url; prompt is optional. resemblance and detail are separate controls, each exposed from 1 to 100 in Raywake.
- Change one control at a time. Record expand_prompt and seed, and request a quote for the selected settings before submitting.

Review the result:
Compare the source and output at equal display size, then inspect the export at full size. Look for changed lettering, invented texture and halos; added detail may not be accurate.

## Ideogram V2
Source: https://raywake.si/explore/ideogram/v2
Model ID: ideogram/v2
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v2/api
Content updated: 2026-10-01

Lay out a poster by stating the headline, subject and space around the text.

Inputs and settings:
- Use prompt, aspect_ratio and style. expand_prompt can elaborate the brief; negative_prompt describes unwanted content.
- Edit mode needs image_url and mask_url. Remix needs image_url and offers strength; these inputs belong to their separate modes.

Review the result:
Proofread every word and inspect letter spacing at the final display size.

## Ideogram V2 Turbo
Source: https://raywake.si/explore/ideogram/v2/turbo
Model ID: ideogram/v2/turbo
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v2/turbo/api
Content updated: 2026-10-01

Explore headline placement and visual hierarchy in a concise poster brief.

Inputs and settings:
- Choose aspect_ratio and style; use negative_prompt for exclusions. Keep seed fixed while comparing expand_prompt.
- Turbo edit requires image_url and mask_url; Turbo remix requires image_url and has strength. The default route creates images from text.

Review the result:
Check misspelled text, crowded margins and unintended objects before reuse.

## Ideogram V2A
Source: https://raywake.si/explore/ideogram/v2a
Model ID: ideogram/v2a
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v2a/api
Content updated: 2026-10-01

Describe the headline independently from the scene so its wording is unambiguous.

Inputs and settings:
- Use prompt and select aspect_ratio and style. Compare expand_prompt with a saved seed.
- To reinterpret a source, use remix with image_url and strength rather than adding images to text mode.

Review the result:
Check the exact wording, text placement and contrast against the background.

## Ideogram V2A Turbo
Source: https://raywake.si/explore/ideogram/v2a/turbo
Model ID: ideogram/v2a/turbo
Provider documentation: https://fal.ai/models/fal-ai/ideogram/v2a/turbo/api
Content updated: 2026-10-01

Choose a style direction, then test a short headline within the composition.

Inputs and settings:
- Text mode accepts prompt, aspect_ratio and style. expand_prompt controls brief expansion; save seed for comparisons.
- Turbo remix adds image_url and strength from 0.01–1. This row does not expose a masked editing mode.

Review the result:
Inspect broken letterforms and whether decorative elements overwhelm the message.

## Ideogram Character
Source: https://raywake.si/explore/ideogram/character
Model ID: ideogram/character
Provider documentation: https://fal.ai/models/fal-ai/ideogram/character/api
Content updated: 2026-10-01

Place a reference character into a described scene with Ideogram Character. Describe the pose and surroundings without relying on the reference to explain the whole scene.

Inputs and settings:
- Provide prompt and one character image in reference_image_urls. The provider uses only the first character reference; image_urls supplies separate style references.
- Choose style or style_codes rather than combining them. Record expand_prompt and rendering_speed, and begin with num_images set to 1.

Review the result:
Compare facial features, clothing and proportions with the reference. Character resemblance is a result to inspect, not a guarantee.

## Ideogram V4.5 Edit
Source: https://raywake.si/explore/ideogram/v4.5/edit
Model ID: ideogram/v4.5/edit
Provider documentation: https://fal.ai/models/ideogram/v4.5/edit/api
Content updated: 2026-10-01

Revise an existing image with Ideogram 4.5 Edit. Describe the intended change and the parts that should remain recognizable.

Inputs and settings:
- Provide image_url and prompt. An optional mask must match the source dimensions: black marks editing, white marks preservation.
- Use at most four reference_image_urls, or three with a mask. Choose quality and num_images, then quote the complete request.

Review the result:
Compare the edited region with the source, including text, shadows and nearby edges.

## Illusion Diffusion
Source: https://raywake.si/explore/illusion-diffusion
Model ID: illusion-diffusion
Provider documentation: https://fal.ai/models/fal-ai/illusion-diffusion/api
Content updated: 2026-10-01

Build a scene around a reference pattern with Illusion Diffusion. Choose a readable pattern before describing the scene that should contain it.

Inputs and settings:
- Send image_url and prompt; controlnet_conditioning_scale controls the reference influence.
- Use control_guidance_start and control_guidance_end to set the conditioning interval; keep seed fixed when comparing.

Review the result:
Check whether both the pattern and the surrounding scene remain readable at the intended display size.

## Image Editing Tools
Source: https://raywake.si/explore/image-editing/age-progression
Model ID: image-editing/age-progression
Provider documentation: https://fal.ai/models/fal-ai/image-editing/age-progression/api
Content updated: 2026-10-01

Explore apparent aging through the default Age Progression mode of Image Editing Tools. Choose a clear portrait for comparison.

Inputs and settings:
- In age-progression mode, send image_url and use prompt to describe the age change.
- Other editing operations have separate schemas. Keep seed and aspect_ratio consistent when comparing age variations.

Review the result:
Compare facial proportions and identity. Generated aging is not a prediction of a person’s future appearance.

## Depth Anything V2 Preprocessor
Source: https://raywake.si/explore/image-preprocessors/depth-anything/v2
Model ID: image-preprocessors/depth-anything/v2
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/depth-anything/v2/api
Content updated: 2026-10-01

Produce a depth-map image from a photograph with Depth Anything V2. Use the map to inspect the scene’s depth structure before another image workflow.

Inputs and settings:
- Provide image_url; no scene prompt is required.
- Inspect the returned image at full size. Confirm that a downstream operation accepts depth maps before passing the result to it.

Review the result:
Compare subject boundaries and depth transitions with the source. Do not treat this image as a calibrated distance measurement.

## HED Edge Preprocessor
Source: https://raywake.si/explore/image-preprocessors/hed
Model ID: image-preprocessors/hed
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/hed/api
Content updated: 2026-10-01

Extract a line map from an image with the HED detector. Use the result to inspect contours before a compatible conditioning workflow.

Inputs and settings:
- Provide image_url. safe selects the detector’s safe variant; scribble selects its scribble variant.
- Change one variant at a time and compare the maps. These controls process the source; they do not describe a new scene.

Review the result:
Check that important outlines survive and background texture does not overwhelm the subject.

## Lineart Preprocessor
Source: https://raywake.si/explore/image-preprocessors/lineart
Model ID: image-preprocessors/lineart
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/lineart/api
Content updated: 2026-10-01

Create a line-art map from a source image. Use it to review which outlines remain available for a subsequent image workflow.

Inputs and settings:
- Send image_url. coarse selects the coarse model variant.
- Compare coarse enabled and disabled on the same source. Keep the chosen map with its source so downstream changes are easier to evaluate.

Review the result:
Inspect small features, broken contours and excessive background lines. A line map needs review before it becomes a design asset.

## MLSD Line Preprocessor
Source: https://raywake.si/explore/image-preprocessors/mlsd
Model ID: image-preprocessors/mlsd
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/mlsd/api
Content updated: 2026-10-01

Detect line segments in an image with MLSD. Start with a source whose important structural lines are clearly visible.

Inputs and settings:
- Provide image_url. score_threshold and distance_threshold both default to 0.1 in the provider schema.
- Adjust one threshold at a time and record it with the output. Compare the line map before using it in another operation.

Review the result:
Compare detected lines with the source. Check missing segments and false lines from texture; this result is not a measured architectural drawing.

## PiDi Edge Preprocessor
Source: https://raywake.si/explore/image-preprocessors/pidi
Model ID: image-preprocessors/pidi
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/pidi/api
Content updated: 2026-10-01

Extract contours with the PiDi image preprocessor. Use the line map to review which features a compatible image workflow could preserve.

Inputs and settings:
- Provide image_url. safe and scribble select detector variants; apply_filter enables its filtering option.
- Change one option at a time and save it with the map. Confirm that the next operation accepts this kind of conditioning image.

Review the result:
Check weak contours, background texture and small subject details. Compare the map with the original at the same size.

## SAM Segmentation Preprocessor
Source: https://raywake.si/explore/image-preprocessors/sam
Model ID: image-preprocessors/sam
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/sam/api
Content updated: 2026-10-01

Produce a segmentation-map image with the SAM preprocessor. Use the map to inspect how the source is divided into regions.

Inputs and settings:
- Send image_url; this preprocessor exposes no text query or selection-point control.
- Review the returned image and the next operation’s requirements. Convert or select regions only when the downstream workflow calls for that representation.

Review the result:
Check adjacent objects and small regions against the original. Do not assume a colored region map is already a transparent foreground cutout.

## Scribble Preprocessor
Source: https://raywake.si/explore/image-preprocessors/scribble
Model ID: image-preprocessors/scribble
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/scribble/api
Content updated: 2026-10-01

Create a scribble-style guide from an existing image. Compare detector variants to see which retains the outlines important to your composition.

Inputs and settings:
- Send image_url and choose model as HED or PiDi. safe selects the detector’s safe variant.
- Keep the source fixed while changing the detector. Review the map before passing it to an operation that accepts scribble conditioning.

Review the result:
Check that the subject remains recognizable and unnecessary background marks do not dominate.

## TeeD Edge Preprocessor
Source: https://raywake.si/explore/image-preprocessors/teed
Model ID: image-preprocessors/teed
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/teed/api
Content updated: 2026-10-01

Generate a TeeD line map for a source photograph. Inspect its contours before choosing it as a guide for another image operation.

Inputs and settings:
- Provide image_url; TeeD exposes no prompt or threshold control in this Raywake operation.
- Check the returned image at full size and confirm downstream compatibility. The map describes source structure rather than generating a new scene.

Review the result:
Review thin edges, texture noise and broken lines. Compare with another preprocessor on the same source if the outlines are unsuitable.

## Zoe Depth Preprocessor
Source: https://raywake.si/explore/image-preprocessors/zoe
Model ID: image-preprocessors/zoe
Provider documentation: https://fal.ai/models/fal-ai/image-preprocessors/zoe/api
Content updated: 2026-10-01

Estimate an image’s depth structure with the Zoe preprocessor. Use its depth-map output when planning a compatible conditioning workflow.

Inputs and settings:
- Send image_url without a prompt. This operation returns a depth-map image.
- Inspect the map and keep it alongside the source. Verify the next operation’s accepted depth representation before passing it onward.

Review the result:
Compare foreground boundaries and depth transitions with the source. Do not infer calibrated real-world distances from the returned picture.

## Image2Pixel
Source: https://raywake.si/explore/image2pixel
Model ID: image2pixel
Provider documentation: https://fal.ai/models/fal-ai/image2pixel/api
Content updated: 2026-10-01

Clean up a pixel-art image with Image2Pixel. Use its palette and grid controls to prepare a consistent sprite or illustration.

Inputs and settings:
- Send image_url; choose max_colors or a fixed_palette for color control.
- Use scale or detection settings for pixel size. Compare the returned processed and scaled images at their own dimensions.

Review the result:
Check isolated pixels, silhouette readability and transparency at native size rather than a smoothed browser zoom.

## Infinity Star
Source: https://raywake.si/explore/infinity-star/text-to-video
Model ID: infinity-star/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/infinity-star/text-to-video/api
Content updated: 2026-10-01

Describe one subject action and a restrained camera movement.

Inputs and settings:
- Use prompt and aspect_ratio; negative_prompt identifies unwanted elements. enhance_prompt can expand the scene description.
- Start with documented guidance_scale 7.5 and tau_video 0.4. Keep use_apg and seed stable while comparing changes; no duration field is exposed.

Review the result:
Check motion consistency, unstable edges and objects that appear between frames.

## Inpainting sdxl and sd
Source: https://raywake.si/explore/inpaint
Model ID: inpaint
Provider documentation: https://fal.ai/models/fal-ai/inpaint/api
Content updated: 2026-10-01

Replace a masked part of an image with the Inpainting route. Prepare a mask aligned with the source before writing the replacement brief.

Inputs and settings:
- Supply image_url, mask_url, model_name and prompt. White mask areas are edited; black areas are preserved.
- Choose a compatible base model through model_name. Keep source and mask dimensions aligned; inspect guidance and seed separately.

Review the result:
Check the edited boundary and surrounding texture for seams or unintended changes.

## Juggernaut Flux Base
Source: https://raywake.si/explore/rundiffusion/juggernaut-flux/base
Model ID: rundiffusion/juggernaut-flux/base
Provider documentation: https://fal.ai/models/rundiffusion-fal/juggernaut-flux/base/api
Content updated: 2026-10-01

Use a clear focal subject and describe the light, textures and camera angle.

Inputs and settings:
- The default requires prompt. Begin with num_images=1; the supported range is 1–4. num_inference_steps accepts 1–50 and guidance_scale accepts 1–20.
- Choose jpeg or png through output_format. Save seed and change one setting at a time to compare the resulting composition.

Review the result:
Inspect texture boundaries, anatomy and any requested letters at full size.

## Juggernaut Flux Lightning
Source: https://raywake.si/explore/rundiffusion/juggernaut-flux/lightning
Model ID: rundiffusion/juggernaut-flux/lightning
Provider documentation: https://fal.ai/models/rundiffusion-fal/juggernaut-flux/lightning/api
Content updated: 2026-10-01

Explore lighting and composition for a portrait or product concept.

Inputs and settings:
- Describe subject, light and framing in prompt. Begin with one image; num_images supports 1–4 outputs.
- The provider starts at four inference steps; Raywake accepts 1–12. Keep seed fixed while revising the brief. Choose JPEG or PNG with output_format.

Review the result:
Inspect skin texture, object edges and lettering at full size.

## Juggernaut Flux Pro
Source: https://raywake.si/explore/rundiffusion/juggernaut-flux/pro
Model ID: rundiffusion/juggernaut-flux/pro
Provider documentation: https://fal.ai/models/rundiffusion-fal/juggernaut-flux/pro/api
Content updated: 2026-10-01

Develop a photographic concept with explicit materials, lighting and background.

Inputs and settings:
- Use prompt to describe the scene; this operation does not take a source photo. num_images allows 1–4 outputs.
- The provider starts at 28 steps and guidance_scale 3.5. Adjust one control at a time within Raywake’s ranges; save seed and choose JPEG or PNG with output_format.

Review the result:
Check reflections, hands and fine textures before using the image.

## Kling 1.0
Source: https://raywake.si/explore/kling-video/v1/standard/effects
Model ID: kling-video/v1/standard/effects
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1/standard/effects/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose an effect and a 5 or 10 second duration. heart_gesture requires exactly two input images; every other effect takes one.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling 1.5
Source: https://raywake.si/explore/kling-video/v1.5/pro/effects
Model ID: kling-video/v1.5/pro/effects
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1.5/pro/effects/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose an effect and a 5 or 10 second duration. heart_gesture requires exactly two input images; every other effect takes one.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling 1.6
Source: https://raywake.si/explore/kling-video/v1.6/pro/effects
Model ID: kling-video/v1.6/pro/effects
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1.6/pro/effects/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose an effect and a 5 or 10 second duration. heart_gesture requires exactly two input images; every other effect takes one.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling 1.6
Source: https://raywake.si/explore/kling-video/v1.6/standard/effects
Model ID: kling-video/v1.6/standard/effects
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1.6/standard/effects/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose an effect and a 5 or 10 second duration. heart_gesture requires exactly two input images; every other effect takes one.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling AI Avatar
Source: https://raywake.si/explore/kling-video/v1/standard/ai-avatar
Model ID: kling-video/v1/standard/ai-avatar
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1/standard/ai-avatar/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Upload an avatar image and an audio file. The credit estimate uses the measured length of your uploaded audio.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling AI Avatar Pro
Source: https://raywake.si/explore/kling-video/v1/pro/ai-avatar
Model ID: kling-video/v1/pro/ai-avatar
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1/pro/ai-avatar/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Upload an avatar image and an audio file. The credit estimate uses the measured length of your uploaded audio.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling AI Avatar v2
Source: https://raywake.si/explore/kling-video/ai-avatar/v2/standard
Model ID: kling-video/ai-avatar/v2/standard
Provider documentation: https://fal.ai/models/fal-ai/kling-video/ai-avatar/v2/standard/api
Content updated: 2026-10-01

Pair a clear character image with finished narration. The audio supplies the speech; the optional prompt describes the visual performance.

Inputs and settings:
- Provide both image_url and audio_url. Listen to the recording beforehand so unwanted pauses or background sound are not part of your input.
- Use prompt for the intended expression or movement. Standard is the default; Pro is a separate operation with the same three exposed fields.

Review the result:
Watch lip movements against the recording. Check facial stability, pauses and gestures throughout the clip.

## Kling O3 Image
Source: https://raywake.si/explore/kling-image/o3/text-to-image
Model ID: kling-image/o3/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/kling-image/o3/text-to-image/api
Content updated: 2026-10-02

Describe the scene and keep subjects consistent with character or object elements. Switch to image-to-image when supplying source images.

Inputs and settings:
- Choose 1K, 2K or 4K resolution. result_type single uses num_images from 1–9; series uses series_amount from 2–9. If the series amount is omitted, the quote reserves up to nine images.
- Elements support a frontal image and, when using extra reference views, 1 or 3 views (2 are not accepted), referenced as @Element1. Image-to-image requires image_urls and accepts up to ten source images.

Review the result:
Inspect details, text and consistency across a series at full size.

## Kling V3 Image
Source: https://raywake.si/explore/kling-image/v3/text-to-image
Model ID: kling-image/v3/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/kling-image/v3/text-to-image/api
Content updated: 2026-10-01

Describe the subject, light and framing for a product concept or illustration. Switch to image-to-image when an existing picture must guide the result.

Inputs and settings:
- Use prompt, then choose aspect_ratio and 1K or 2K resolution. Start with one image through num_images; the supported range is 1–9.
- Select JPEG, PNG or WebP with output_format. Image-to-image additionally requires image_url and prompt; specify the changes you want.

Review the result:
Inspect lettering, faces and repeated objects at full size; compare edited details with the source.

## Kling O1 Edit Video [Pro]
Source: https://raywake.si/explore/kling-video/o1/video-to-video/edit
Model ID: kling-video/o1/video-to-video/edit
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o1/video-to-video/edit/api
Content updated: 2026-10-01

Use an existing clip when you want to reinterpret its appearance. Separate the desired change from the movement and framing you want to keep.

Inputs and settings:
- Provide video_url and prompt. The documented source is MP4/MOV, 3–10 seconds and no larger than 200 MB; trim a longer clip before uploading.
- Use keep_audio for the original soundtrack. Keep the source and instruction together when comparing edits; this operation does not expose a new duration setting.

Review the result:
Compare the subject across frames, especially hands, clothing and overlap boundaries. Check retained sound separately.

## Kling O1 Edit Video [Standard]
Source: https://raywake.si/explore/kling-video/o1/standard/video-to-video/edit
Model ID: kling-video/o1/standard/video-to-video/edit
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o1/standard/video-to-video/edit/api
Content updated: 2026-10-01

Describe a focused change to the setting, subject or visual style, and name the motion that should remain recognizable.

Inputs and settings:
- Supply video_url and a required prompt, up to 2500 characters. Official input guidance specifies MP4/MOV clips of 3–10 seconds, up to 200 MB.
- Set keep_audio deliberately. Begin with one edit so you can assess unintended changes before revising the brief.

Review the result:
Compare movement, camera framing and edited edges with the original; listen if source audio was retained.

## Kling O1
Source: https://raywake.si/explore/kling-video/o1/image-to-video
Model ID: kling-video/o1/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o1/image-to-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose 3–10 seconds. For an element, upload a frontal image plus 1–3 extra views and reference it with #Element1 in the prompt.
- With a source video, use at most four elements and separate reference images combined. Other reference modes allow seven. Source videos must be MP4 or MOV, 3–10 seconds, 720–2160 pixels, 24–60 fps and at most 200 MB.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling O1 Standard
Source: https://raywake.si/explore/kling-video/o1/standard/image-to-video
Model ID: kling-video/o1/standard/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o1/standard/image-to-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose 3–10 seconds. For an element, upload a frontal image plus 1–3 extra views and reference it with #Element1 in the prompt.
- With a source video, use at most four elements and separate reference images combined. Other reference modes allow seven. Source videos must be MP4 or MOV, 3–10 seconds, 720–2160 pixels, 24–60 fps and at most 200 MB.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Kling O1 Image
Source: https://raywake.si/explore/kling-image/o1
Model ID: kling-image/o1
Provider documentation: https://fal.ai/models/fal-ai/kling-image/o1/api
Content updated: 2026-10-01

Assign each reference a clear role, such as subject, clothing or background.

Inputs and settings:
- Send image_urls and prompt; refer to uploaded images as @Image1, @Image2 and so on in their order.
- Choose 1K or 2K resolution, aspect_ratio and num_images from 1–9. Describe what must remain unchanged as well as the edit.

Review the result:
Compare identity, object proportions and boundaries where references meet.

## Kling O3 Pro
Source: https://raywake.si/explore/kling-video/o3/pro/text-to-video
Model ID: kling-video/o3/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o3/pro/text-to-video/api
Content updated: 2026-10-02

Choose text-to-video, image-to-video, reference-to-video, video editing or video-to-video reference for the source you have.

Inputs and settings:
- For text or image generation, use prompt or multi_prompt, not both. Choose 3–15 seconds; generate_audio controls generated sound. Image mode requires image_url.
- Reference and Edit modes support elements: add a frontal image and 1 or 3 extra views per image element (2 extra views are not accepted), then use @Element1 in the prompt. image_urls are separate references, named @Image1.
- Edit requires video_url and prompt. keep_audio controls source sound. A video input permits at most four elements and separate reference images combined; the extra views inside an element stay grouped with it.

Review the result:
Review character identity, motion, camera movement and the soundtrack throughout the clip. Compare edits with the source.

## Kling O3 Standard
Source: https://raywake.si/explore/kling-video/o3/standard/text-to-video
Model ID: kling-video/o3/standard/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o3/standard/text-to-video/api
Content updated: 2026-10-02

Choose text-to-video, image-to-video, reference-to-video, video editing or video-to-video reference for the source you have.

Inputs and settings:
- For text or image generation, use prompt or multi_prompt, not both. Choose 3–15 seconds; generate_audio controls generated sound. Image mode requires image_url.
- Reference and Edit modes support elements: add a frontal image and 1 or 3 extra views per image element (2 extra views are not accepted), then use @Element1 in the prompt. image_urls are separate references, named @Image1.
- Edit requires video_url and prompt. keep_audio controls source sound. A video input permits at most four elements and separate reference images combined; the extra views inside an element stay grouped with it.

Review the result:
Review character identity, motion, camera movement and the soundtrack throughout the clip. Compare edits with the source.

## Kling TTS
Source: https://raywake.si/explore/kling-video/v1/tts
Model ID: kling-video/v1/tts
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v1/tts/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## Kling O3 4K
Source: https://raywake.si/explore/kling-video/o3/4k/text-to-video
Model ID: kling-video/o3/4k/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/o3/4k/text-to-video/api
Content updated: 2026-10-02

Choose text-to-video, image-to-video, reference-to-video, video editing or video-to-video reference for the source you have.

Inputs and settings:
- For text or image generation, use prompt or multi_prompt, not both. Choose 3–15 seconds; generate_audio controls generated sound. Image mode requires image_url.
- Reference and Edit modes support elements: add a frontal image and 1 or 3 extra views per image element (2 extra views are not accepted), then use @Element1 in the prompt. image_urls are separate references, named @Image1.
- Edit requires video_url and prompt. keep_audio controls source sound. A video input permits at most four elements and separate reference images combined; the extra views inside an element stay grouped with it.

Review the result:
Review character identity, motion, camera movement and the soundtrack throughout the clip. Compare edits with the source.

## Kling 2.5 Turbo Standard
Source: https://raywake.si/explore/kling-video/v2.5-turbo/standard/image-to-video
Model ID: kling-video/v2.5-turbo/standard/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v2.5-turbo/standard/image-to-video/api
Content updated: 2026-10-01

Choose a still with a clear subject, then describe its action and the camera movement. Avoid asking for a wholly unrelated composition.

Inputs and settings:
- Both image_url and prompt are required. Select duration of 5 or 10 seconds; the current operation has no separate aspect_ratio control.
- Adjust cfg_scale within 0–1 when testing prompt adherence. negative_prompt describes unwanted details; keep other inputs fixed when comparing a change.

Review the result:
Compare the opening with your image and inspect small moving details. Check for background warping around the subject.

## Kling Video to Audio
Source: https://raywake.si/explore/kling-video/video-to-audio
Model ID: kling-video/video-to-audio
Provider documentation: https://fal.ai/models/fal-ai/kling-video/video-to-audio/api
Content updated: 2026-10-01

Describe scene-specific sound effects and music separately. This operation supplies sound for a clip rather than reading a narration script.

Inputs and settings:
- Provide video_url. Official input guidance specifies MP4/MOV, 3–20 seconds and a file no larger than 100 MB.
- sound_effect_prompt and background_music_prompt each allow 200 characters. Use asmr_mode when seeking detailed effects, then review the actual mix.

Review the result:
Listen for alignment with visible actions and sudden level changes. Check whether music masks important effects.

## Kling 3 Turbo Standard
Source: https://raywake.si/explore/kling-video/v3/turbo/standard/text-to-video
Model ID: kling-video/v3/turbo/standard/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v3/turbo/standard/text-to-video/api
Content updated: 2026-10-01

Build a short scene around a specific subject and movement. Choose the image operation if you already have the composition as a still.

Inputs and settings:
- Use prompt for text generation, duration of 3–15 seconds and aspect_ratio of 16:9, 9:16 or 1:1.
- The image operation needs image_url; its prompt describes movement within that source. Official guidance recommends prompts under 2500 characters, despite the larger Raywake field limit.

Review the result:
Watch for changing shapes, background drift and abrupt final movement. Check the clip in the layout where it will be used.

## Kling 3 4K
Source: https://raywake.si/explore/kling-video/v3/4k/text-to-video
Model ID: kling-video/v3/4k/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v3/4k/text-to-video/api
Content updated: 2026-10-01

Use a compact visual brief for a new scene, keeping the action readable. Select image-to-video when a specific opening image matters.

Inputs and settings:
- Write prompt, then select duration of 3–15 seconds and aspect_ratio. Choose generate_audio; cfg_scale within 0–1 adjusts prompt adherence.
- Use negative_prompt for unwanted details. Image-to-video requires start_image_url and optionally accepts end_image_url; choose compatible frames before describing their transition.

Review the result:
Inspect fine moving detail at full size, particularly faces and hard edges. Review generated speech and the closing frame.

## Kling 3 Turbo Pro
Source: https://raywake.si/explore/kling-video/v3/turbo/pro/text-to-video
Model ID: kling-video/v3/turbo/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v3/turbo/pro/text-to-video/api
Content updated: 2026-10-01

Keep one main action and a clear camera direction in the brief. Use image-to-video when an existing still should define the opening.

Inputs and settings:
- Write prompt and choose duration from 3–15 seconds. Set aspect_ratio to 16:9, 9:16 or 1:1 for text generation.
- Image-to-video requires image_url and has its own duration and prompt fields. Keep the brief concise; official guidance recommends fewer than 2500 characters.

Review the result:
Check whether the subject stays recognizable throughout the action. Review the opening and ending for a useful edit point.

## Kling Video v2.6 Motion Control [Standard]
Source: https://raywake.si/explore/kling-video/v2.6/standard/motion-control
Model ID: kling-video/v2.6/standard/motion-control
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v2.6/standard/motion-control/api
Content updated: 2026-10-01

Pair an unobstructed character image with a reference video showing the intended body movement. Keep the head and body readable in both.

Inputs and settings:
- Provide image_url, video_url and required character_orientation. Official guidance allows up to 10 seconds for image orientation or 30 seconds for video orientation.
- Use keep_original_sound to retain reference audio. An optional prompt can describe scene details; motion still comes from the reference video.

Review the result:
Compare the pose sequence, hands and body proportions against the reference. Listen to retained sound if enabled.

## Kling 2.6 Pro
Source: https://raywake.si/explore/kling-video/v2.6/pro/text-to-video
Model ID: kling-video/v2.6/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v2.6/pro/text-to-video/api
Content updated: 2026-10-01

Describe action and ambience for a new clip. Use image-to-video for an opening still, or motion control to follow a reference performance.

Inputs and settings:
- Default text generation requires prompt. Select duration of 5 or 10 seconds, aspect_ratio and generate_audio; cfg_scale controls prompt adherence.
- Native voice output is documented for English and Chinese; other languages may be translated to English. Image-to-video uses start_image_url; motion control needs image_url, video_url and character_orientation.

Review the result:
Watch movement and listen to speech separately. Check pronunciation, lip timing and whether ambience overwhelms the voice.

## Kling 3 Pro
Source: https://raywake.si/explore/kling-3-pro
Model ID: kling-video/v3/pro/text-to-video
Content updated: 2026-10-01

Start with a strong still frame when the composition matters. The coastal-drive sample shows a moving subject within an established landscape. Use an image you have permission to upload, then describe the movement you want to add.

Suggested starting prompt (not the exact prompt behind the preview):
A green sports car follows the coastal road. The camera tracks smoothly beside it while warm evening light moves across the bodywork. Keep the scene continuous.

Inputs and settings:
- To animate your photo, select the image-to-video mode and provide start_image_url. The default text-to-video mode starts from a prompt instead. For a source photo, describe movement within its existing composition.
- For one shot, use prompt; multi_prompt is for multiple shots. Choose one approach rather than sending both. Select duration from the current schema and quote the complete request.
- Use generate_audio to decide whether to include sound. If using end_image_url, choose a compatible ending frame and check whether the transition remains plausible.

Review the result:
Watch the whole clip. Check that the subject stays consistent, the camera follows your direction and small details remain plausible as they move.

### Do I need a Kling subscription to use Kling 3 Pro on Raywake?
Raywake uses your Raywake credit balance for each generation. Choose the operation and full input, then review its quote. A Kling subscription price or daily credit allowance belongs to Kling’s own service and does not set your Raywake balance or generation price.
Raywake credits: https://raywake.si/docs/credits

## Kling 3 Standard
Source: https://raywake.si/explore/kling-video/v3/standard/text-to-video
Model ID: kling-video/v3/standard/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v3/standard/text-to-video/api
Content updated: 2026-10-01

Describe one main action and camera movement. Use image-to-video for a still, or motion control when another video should guide the character.

Inputs and settings:
- Supply prompt for the default operation. Choose duration of 3–15 seconds, aspect_ratio and generate_audio; cfg_scale adjusts prompt adherence within 0–1.
- Image-to-video requires start_image_url. Motion control instead requires image_url, video_url and character_orientation, with keep_original_sound for reference audio.

Review the result:
Inspect the action from beginning to end. Check changing faces, body proportions and any generated dialogue.

## Kling 2.5 Turbo Pro
Source: https://raywake.si/explore/kling-video/v2.5-turbo/pro/text-to-video
Model ID: kling-video/v2.5-turbo/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/kling-video/v2.5-turbo/pro/text-to-video/api
Content updated: 2026-10-01

Write one clear action for a new scene. If an existing picture should set the composition, switch to the image-to-video operation.

Inputs and settings:
- The default requires prompt; select duration of 5 or 10 seconds and aspect_ratio. cfg_scale ranges from 0–1 and adjusts adherence to your prompt.
- Use negative_prompt for unwanted details. Image-to-video requires image_url and prompt; its optional ending-image field is tail_image_url.

Review the result:
Review faces, hands and camera movement through the full clip. Check that the requested action reaches a useful ending.

## Kokoro TTS
Source: https://raywake.si/explore/kokoro/american-english
Model ID: kokoro/american-english
Provider documentation: https://fal.ai/models/fal-ai/kokoro/american-english/api
Content updated: 2026-10-01

Use finished narration text rather than a description of a voice. Punctuation helps you plan pauses before generating the recording.

Inputs and settings:
- Provide nonempty text in prompt and choose a documented American English voice ID through voice, such as af_heart.
- Start speed at 1, then adjust only pacing. Raywake accepts 0.1–5; extreme values need listening checks rather than assuming speech stays intelligible.

Review the result:
Listen to names, numbers and abbreviations against the script. Check pacing at the intended playback speed.

## Kolors
Source: https://raywake.si/explore/kolors
Model ID: kolors
Provider documentation: https://fal.ai/models/fal-ai/kolors/api
Content updated: 2026-10-01

Describe the composition, lighting and subject for a new image. Choose image-to-image when you want to reinterpret an existing picture.

Inputs and settings:
- Provide prompt. Adjust guidance_scale within 1–10 or num_inference_steps within 1–150 one at a time; preserve seed when comparing.
- Begin with num_images set to 1 and select JPEG or PNG. Image-to-image requires image_url and prompt, with strength controlling the degree of change.

Review the result:
Inspect anatomy, lettering and small repeated details. For image-to-image, compare how much the source changed.

## Krea 2 Large
Source: https://raywake.si/explore/krea/v2/large/text-to-image
Model ID: krea/v2/large/text-to-image
Provider documentation: https://fal.ai/models/krea/v2/large/text-to-image/api
Content updated: 2026-10-01

Describe the scene, then add supported Krea styles or a moodboard when you need a shared visual direction.

Inputs and settings:
- Provide prompt and choose aspect_ratio. creativity controls how freely the model interprets the brief; keep seed fixed when comparing this setting.
- styles accepts up to 10 entries and moodboards up to one. They reference Krea catalog IDs; an arbitrary uploaded image URL is not a style ID.

Review the result:
Check whether creativity changed important objects or layout. Inspect lettering and proportions at full size.

## Krea 2 Medium
Source: https://raywake.si/explore/krea/v2/medium/text-to-image
Model ID: krea/v2/medium/text-to-image
Provider documentation: https://fal.ai/models/krea/v2/medium/text-to-image/api
Content updated: 2026-10-01

Explore a scene with a specified palette and composition. Add catalog styles or a moodboard only when they support that brief.

Inputs and settings:
- Use required prompt and select aspect_ratio. Adjust creativity to explore interpretation; save seed and the full brief when comparing variations.
- Use styles for up to 10 Krea style entries or moodboards for one moodboard. Both need catalog IDs, not file URLs; verify their nested fields in the studio.

Review the result:
Check subject placement and whether styling obscured required details. Compare results at the same displayed size.

## Krea 2 Medium Turbo
Source: https://raywake.si/explore/krea/v2/medium/turbo/text-to-image
Model ID: krea/v2/medium/turbo/text-to-image
Provider documentation: https://fal.ai/models/krea/v2/medium/turbo/text-to-image/api
Content updated: 2026-10-01

Use a stable brief to compare this variant with other Krea operations. Keep the same subject, framing and style choices for a useful comparison.

Inputs and settings:
- Provide prompt and choose aspect_ratio. Record creativity and seed; change one setting at a time when testing a different interpretation.
- styles supports up to 10 Krea catalog entries; moodboards supports one catalog moodboard. These are distinct from uploaded reference photos and require their proper IDs.

Review the result:
Inspect small textures, object edges and requested lettering. Judge the output against your brief rather than the variant name.

## Krea 2 Turbo
Source: https://raywake.si/explore/krea-2/turbo
Model ID: krea-2/turbo
Provider documentation: https://fal.ai/models/fal-ai/krea-2/turbo/api
Content updated: 2026-10-01

Use a concrete visual brief for image variations. Keep prompt expansion unchanged when comparing the effect of your own wording.

Inputs and settings:
- Provide prompt, then set enable_prompt_expansion deliberately. Save seed and acceleration with the brief; changing multiple controls makes comparison harder.
- Use num_images from 1–4 and choose JPEG or PNG through output_format. This operation exposes no source-image or aspect-ratio input in Raywake.

Review the result:
Compare subject placement, fine lettering and textures across outputs. Inspect the full-size image before choosing it.

## LTX 2.3 Fast
Source: https://raywake.si/explore/ltx-2.3/text-to-video/fast
Model ID: ltx-2.3/text-to-video/fast
Provider documentation: https://fal.ai/models/fal-ai/ltx-2.3/text-to-video/fast/api
Content updated: 2026-10-01

Plan a continuous shot from a scene description with LTX 2.3 Fast. Give the subject one action that fits the duration.

Inputs and settings:
- This text-to-video operation requires prompt. Choose aspect_ratio, duration, resolution and fps; use the separate image-to-video mode for a source photo.
- Durations above 10 seconds require 25 FPS and 1080p. Choose a supported combination before requesting a quote, rather than treating duration, frame rate and resolution as independent.

Review the result:
Watch for object continuity, camera jumps and distorted movement. Review sound when present and record the actual duration.

## LTX 2.3 Pro
Source: https://raywake.si/explore/ltx-2.3/text-to-video
Model ID: ltx-2.3/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/ltx-2.3/text-to-video/api
Content updated: 2026-10-01

Describe the action, framing and sound for one shot. The image operation can connect a starting image to an optional ending image.

Inputs and settings:
- prompt is required. Choose duration of 6, 8 or 10 seconds, aspect_ratio of 16:9 or 9:16, resolution and fps.
- Image-to-video adds image_url, end_image_url and auto aspect ratio. Raywake has no audio-generation switch on this Pro route.

Review the result:
Review movement and audio together; check transitions when supplying an ending frame.

## LTX-2.3 22B
Source: https://raywake.si/explore/ltx-2.3-22b/text-to-video
Model ID: ltx-2.3-22b/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/ltx-2.3-22b/text-to-video/api
Content updated: 2026-10-01

Use this route to compare audio and video guidance for a single scene. Describe each sound alongside its visible cause.

Inputs and settings:
- Provide prompt; choose num_frames and generate_audio. Raywake exposes 8–50 inference steps on this route.
- Change video_cfg_scale or audio_cfg_scale separately while holding seed fixed. The image operation adds image_url and image_strength; neither belongs to default text generation.

Review the result:
Check sound timing, repeated textures and object continuity across the entire clip.

## LTX-2.3 22B Distilled
Source: https://raywake.si/explore/ltx-2.3-22b/distilled/text-to-video
Model ID: ltx-2.3-22b/distilled/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/ltx-2.3-22b/distilled/text-to-video/api
Content updated: 2026-10-01

Describe a continuous action, camera movement and audible events. Use the image-to-video operation when a starting image matters.

Inputs and settings:
- prompt is required. Control the frame count with num_frames and include or omit generated sound with generate_audio.
- Keep seed fixed while comparing enable_prompt_expansion or use_multiscale. image_strength is available only in the image operation; Raywake exposes no FPS control here.

Review the result:
Watch object continuity and listen for sound that misses the action; inspect the ending before editing.

## LCM Diffusion
Source: https://raywake.si/explore/fast-lcm-diffusion
Model ID: fast-lcm-diffusion
Provider documentation: https://fal.ai/models/fal-ai/fast-lcm-diffusion/api
Content updated: 2026-10-01

Explore broad composition and color ideas before refining small details.

Inputs and settings:
- The official starting point is six steps with guidance_scale 1.5. Raywake supports 1–32 steps and up to eight images.
- Image-to-image requires image_url; inpainting also requires mask_url. Both add strength, while text generation starts from prompt alone.

Review the result:
Inspect flat textures, softened edges and whether prompt expansion added unwanted objects.

## Live Portrait
Source: https://raywake.si/explore/live-portrait
Model ID: live-portrait
Provider documentation: https://fal.ai/models/fal-ai/live-portrait/api
Content updated: 2026-10-01

Choose a clear portrait and a driving clip with readable expressions. The separate image mode adjusts a still face without a driving video.

Inputs and settings:
- The default mode requires image_url and video_url. Crop controls and flag_pasteback affect how the animated face fits the original frame.
- Adjust eye or lip retargeting separately. Change smile, blink or head rotation gradually rather than pushing several expression controls together.

Review the result:
Watch the eyes, mouth and face boundary for distortion; compare identity with the source portrait.

## Longcat Image
Source: https://raywake.si/explore/longcat-image
Model ID: longcat-image
Provider documentation: https://fal.ai/models/fal-ai/longcat-image/api
Content updated: 2026-10-01

Explore compositions that combine a subject, lighting and written labels. Spell out the exact text and where it should appear.

Inputs and settings:
- prompt is required. Raywake offers num_images from 1–4 and JPEG, PNG or WebP through output_format.
- Compare guidance_scale and num_inference_steps one at a time, keeping seed fixed. The current Raywake inputs do not expose a size control.

Review the result:
Inspect lettering character by character and check reflections, hands and repeated objects.

## LTX-2.3 Quality
Source: https://raywake.si/explore/ltx-2.3-quality/text-to-video
Model ID: ltx-2.3-quality/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/ltx-2.3-quality/text-to-video/api
Content updated: 2026-10-01

Plan a shot with a clear action and ending. Use image-to-video for motion built from a source still.

Inputs and settings:
- prompt is required. Raywake exposes num_frames, generate_audio, guidance_scale and 8–30 inference steps.
- Compare guidance_scale and num_inference_steps independently with a fixed seed. image_strength belongs to the image operation; do not borrow resolution controls from other LTX routes.

Review the result:
Inspect faces during movement, background stability and sound transitions.

## Luma Ray 3.2
Source: https://raywake.si/explore/luma/agent/ray/v3.2/text-to-video
Model ID: luma/agent/ray/v3.2/text-to-video
Provider documentation: https://fal.ai/models/luma/agent/ray/v3.2/text-to-video/api
Content updated: 2026-10-01

Plan one scene, or use the separate image/video operations to animate a still or edit footage.

Inputs and settings:
- The default needs prompt. Choose duration of 5s or 10s and resolution of 540p, 720p or 1080p. loop supports 5s text generations.
- Image-to-video adds image_url and end_image_url. Video editing requires video_url; auto_controls cannot be combined with edit_strength.

Review the result:
Check object continuity and framing; replay the boundary when requesting a loop.

## Luma Uni-1
Source: https://raywake.si/explore/luma/agent/uni-1/v1/text-to-image
Model ID: luma/agent/uni-1/v1/text-to-image
Provider documentation: https://fal.ai/models/luma/agent/uni-1/v1/text-to-image/api
Content updated: 2026-10-01

Explore an illustration or photo concept by defining the subject, background and light. Its edit operation works from an existing image.

Inputs and settings:
- Provide prompt and choose aspect_ratio for the intended placement. style offers auto or manga; output_format offers PNG or JPEG.
- Use image_url with prompt in the edit operation. Aspect ratio is a text-generation control in Raywake and is absent from the edit DTO.

Review the result:
Check the composition, exact lettering and relationships between objects.

## Luma Uni-1 Max
Source: https://raywake.si/explore/luma/agent/uni-1/v1/max
Model ID: luma/agent/uni-1/v1/max
Provider documentation: https://fal.ai/models/luma/agent/uni-1/v1/max/api
Content updated: 2026-10-01

Compose a main visual with clear subject placement, lighting and style. Choose the separate edit operation to revise an existing image.

Inputs and settings:
- Text generation requires prompt and offers aspect_ratio, style and output_format. style can be auto or manga; output is PNG or JPEG.
- The edit operation requires image_url and prompt. Current Raywake inputs expose no web search or multiple reference images.

Review the result:
Inspect small details and lettering at full size; judge the actual result before choosing it as a main visual.

## Lumina Image 2
Source: https://raywake.si/explore/lumina-image/v2
Model ID: lumina-image/v2
Provider documentation: https://fal.ai/models/fal-ai/lumina-image/v2/api
Content updated: 2026-10-01

Describe the subject and composition; use negative_prompt for elements you want to avoid.

Inputs and settings:
- prompt is required. Choose num_images from 1–4 and JPEG or PNG using output_format.
- Keep seed fixed to compare guidance_scale. Leave system_prompt and CFG normalization controls unchanged until the basic scene is right; Raywake exposes no image-size field.

Review the result:
Inspect lettering, fine textures and crowded areas rather than assessing only the overall composition.

## Lyria 3 Pro
Source: https://raywake.si/explore/lyria3/pro
Model ID: lyria3/pro
Provider documentation: https://fal.ai/models/fal-ai/lyria3/pro/api
Content updated: 2026-10-01

Describe genre, instruments, pace, vocals and song structure. An optional image can provide visual inspiration.

Inputs and settings:
- prompt is required; image_url is optional. Describe tempo and structure in the prompt rather than expecting dedicated duration or BPM sliders.
- For a backing track, explicitly describe the desired instrumental arrangement. The catalog preview is artwork, not an audio sample.

Review the result:
Listen to the entire track, checking sung words, transitions and the ending before placing it under a video.

## Lyria 3.5
Source: https://raywake.si/explore/google/lyria-3.5
Model ID: google/lyria-3.5
Provider documentation: https://fal.ai/models/google/lyria-3.5/api
Content updated: 2026-10-01

Specify genre, instruments, tempo, vocal character and song sections.

Inputs and settings:
- Use prompt for the music brief. image_url can provide visual inspiration.
- Describe desired length or section timings in the prompt; there is no dedicated duration field.

Review the result:
Listen for distorted vocals, abrupt transitions and an ending that fits your edit.

## Lyria3
Source: https://raywake.si/explore/lyria3
Model ID: lyria3
Content updated: 2026-10-01

Describe a musical idea with a genre, instruments, mood and pace. This endpoint generates audio; its catalog artwork is an illustration, not an audio sample or a quality measurement.

Suggested starting prompt (not the exact prompt behind the preview):
A calm instrumental piece with acoustic guitar, gentle percussion and a warm bass line. Keep the arrangement sparse enough to sit behind spoken narration.

Inputs and settings:
- Start with a prompt describing the arrangement you want. Do not send a speech script expecting a text-to-speech reading.
- Check the current schema and quote before submitting. Save your prompt with the result; changing the musical brief can produce a different arrangement.

Review the result:
Listen through the entire result, including any vocals. Check transitions and the ending before placing it under a video; review the applicable usage terms for your project.

## Marey Realism 1.5
Source: https://raywake.si/explore/moonvalley/marey/t2v
Model ID: moonvalley/marey/t2v
Provider documentation: https://fal.ai/models/moonvalley/marey/t2v/api
Content updated: 2026-10-01

Describe a continuous scene with distinct subject and camera movement.

Inputs and settings:
- The default needs prompt; duration offers 5s or 10s. dimensions controls framing, while negative_prompt lists unwanted features.
- Image-to-video adds image_url. Motion Transfer and Pose Transfer instead require video_url and prompt; their optional first_frame_image_url and reference_image_url anchor appearance.

Review the result:
Review motion continuity and compare any transferred movement with the control video.

## Marigold Depth Estimation
Source: https://raywake.si/explore/imageutils/marigold-depth
Model ID: imageutils/marigold-depth
Provider documentation: https://fal.ai/models/fal-ai/imageutils/marigold-depth/api
Content updated: 2026-10-01

Estimate a depth map from a single photo with Marigold. Use the map as a visual input for a compatible downstream workflow.

Inputs and settings:
- Send image_url. processing_res controls processing resolution; zero uses the input size.
- num_inference_steps and ensemble_size control denoising and prediction averaging. Compare the map before increasing either.

Review the result:
Check foreground separation and occluded edges. Do not interpret the image as measured distances.

## Meta Muse Image
Source: https://raywake.si/explore/meta/muse-image/text-to-image
Model ID: meta/muse-image/text-to-image
Provider documentation: https://fal.ai/models/meta/muse-image/text-to-image/api
Content updated: 2026-10-01

Plan a visual containing text or diagrams by spelling out labels and layout. Use edit for an existing image or reference set.

Inputs and settings:
- Text generation requires prompt. Choose aspect_ratio, num_images and output_format; a ratio specifies shape, not exact pixel dimensions.
- The edit operation requires prompt and image_urls, accepting 1–10 source images in Raywake. Explain the role of each image in the instruction.

Review the result:
Read every label, verify plotted values and scan any QR code before relying on it.

## Midas Depth Estimation
Source: https://raywake.si/explore/imageutils/depth
Model ID: imageutils/depth
Provider documentation: https://fal.ai/models/fal-ai/imageutils/depth/api
Content updated: 2026-10-01

Extract a depth representation from a photo with MiDaS. Keep the source image alongside the returned map for inspection.

Inputs and settings:
- Send image_url; no text prompt is required.
- Use the default a and bg_th values initially. The provider schema gives no detailed interpretation of these controls.

Review the result:
Compare near and far surfaces and object edges. A monocular estimate is not a survey or metric measurement.

## MiniMax (Hailuo AI) Music
Source: https://raywake.si/explore/minimax-music
Model ID: minimax-music
Provider documentation: https://fal.ai/models/fal-ai/minimax-music/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Write the musical direction and lyrics in the fields supplied by this version. Where audio settings are available, choose sample rate, bitrate and format.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## MiniMax (Hailuo AI) Music v1.5
Source: https://raywake.si/explore/minimax-music/v1.5
Model ID: minimax-music/v1.5
Provider documentation: https://fal.ai/models/fal-ai/minimax-music/v1.5/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Write the musical direction and lyrics in the fields supplied by this version. Where audio settings are available, choose sample rate, bitrate and format.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## MiniMax Image-01
Source: https://raywake.si/explore/minimax/image-01
Model ID: minimax/image-01
Provider documentation: https://fal.ai/models/fal-ai/minimax/image-01/api
Content updated: 2026-10-01

Describe the subject and setting, then use a reference when appearance matters.

Inputs and settings:
- The default requires prompt, limited to 1500 characters. Subject-reference also requires image_url; a text-only request does not use your photograph.
- Choose aspect_ratio for the composition and num_images for the output count. prompt_optimizer controls automatic rewriting; change one setting when comparing variations.

Review the result:
Compare facial features and clothing against the reference; inspect hands and lettering.

## MiniMax Video-01
Source: https://raywake.si/explore/minimax/video-01
Model ID: minimax/video-01
Provider documentation: https://fal.ai/models/fal-ai/minimax/video-01/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax Video-01 Director
Source: https://raywake.si/explore/minimax/video-01-director
Model ID: minimax/video-01-director
Provider documentation: https://fal.ai/models/fal-ai/minimax/video-01-director/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax Video-01 Live
Source: https://raywake.si/explore/minimax/video-01-live
Model ID: minimax/video-01-live
Provider documentation: https://fal.ai/models/fal-ai/minimax/video-01-live/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax (Hailuo AI) Video 01 Subject Reference
Source: https://raywake.si/explore/minimax/video-01-subject-reference
Model ID: minimax/video-01-subject-reference
Provider documentation: https://fal.ai/models/fal-ai/minimax/video-01-subject-reference/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax H3 Max
Source: https://raywake.si/explore/minimax/h3-max/text-to-video
Model ID: minimax/h3-max/text-to-video
Content updated: 2026-10-01

Choose the mode for the material you already have. Start with text for a new scene, image-to-video for an opening photo, or a video-editing mode for an existing clip. The default endpoint on this page creates video from text; the other modes have their own model IDs.

Suggested starting prompt (not the exact prompt behind the preview):
A cyclist rides along a quiet riverside path in warm morning light. The camera follows from the side at a steady distance. Keep the rider and bicycle consistent in one continuous shot.

Inputs and settings:
- Describe the subject, one main action and camera movement in prompt. Choose duration, resolution and aspect_ratio from the current schema before requesting a quote.
- prompt_expansion_mode controls whether the prompt is rewritten. When comparing your own instructions, keep the same expansion setting and record any expanded prompt returned with the result.
- An optional target_audio_url can supply a soundtrack. Review the supported file constraints and video duration first; listen for trimming, silence or a mismatch with the visual action.

Review the result:
Watch the whole clip and listen to the soundtrack. Check the bicycle wheels, body movement, background continuity and any requested speech. Keep the original input and selected settings when comparing another mode.

## MiniMax H3
Source: https://raywake.si/explore/minimax/h3/text-to-video
Model ID: minimax/h3/reference-to-video
Provider documentation: https://fal.ai/models/minimax/h3/text-to-video
Content updated: 2026-10-01

Plan the subject’s action, camera movement and sound. Separate image and reference operations work from supplied visuals.

Inputs and settings:
- Raywake requires prompt and exposes duration from 5–15 seconds, aspect_ratio and resolution. Use the current operation’s fields when requesting a quote.
- Image-to-video offers image_url and end_image_url; reference-to-video offers reference_image_urls. Keep these inputs in their corresponding operation.

Review the result:
Watch for changing identities, disappearing objects and sound that misses the action.

## MiniMax H3 LoRA
Source: https://raywake.si/explore/minimax/h3/image-to-video/lora
Model ID: minimax/h3/reference-to-video/lora
Provider documentation: https://fal.ai/models/minimax/h3/image-to-video/lora/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Choose 5–15 seconds and 480P, 768P, 2K or 4K. Add up to three LoRA adapters with their paths and scales.
- Reference mode accepts up to nine images, three videos and three audio clips, with at most 12 files overall. Each video or audio clip must be 2–15 seconds; video durations together and audio durations together must each stay within 15 seconds.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax Hailuo 02 Pro
Source: https://raywake.si/explore/minimax/hailuo-02/pro/text-to-video
Model ID: minimax/hailuo-02/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-02/pro/text-to-video/api
Content updated: 2026-10-01

Describe a short action and camera movement. Choose the separate image operation when you need an opening or ending still.

Inputs and settings:
- The default requires prompt and offers prompt_optimizer. Raywake limits the prompt to 2000 characters.
- Image-to-video requires image_url and prompt, with optional end_image_url. This Pro DTO exposes no duration or resolution selectors.

Review the result:
Check motion continuity and small details; inspect both ends when using a frame transition.

## MiniMax Hailuo 02 Standard
Source: https://raywake.si/explore/minimax/hailuo-02/standard/text-to-video
Model ID: minimax/hailuo-02/standard/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-02/standard/text-to-video/api
Content updated: 2026-10-01

Plan an action that fits the chosen duration. Switch to image-to-video for a starting still or a transition between frames.

Inputs and settings:
- Text generation requires prompt and offers duration of 6 or 10 seconds plus prompt_optimizer. Keep the prompt within 2000 characters.
- Only image-to-video exposes resolution of 512P or 768P, along with image_url and optional end_image_url. These are separate from default text controls.

Review the result:
Review subject appearance, camera movement and the ending; compare supplied frames with the output.

## MiniMax Hailuo 2.3 Fast [Pro] (Image to Video)
Source: https://raywake.si/explore/minimax/hailuo-2.3-fast/pro/image-to-video
Model ID: minimax/hailuo-2.3-fast/pro/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-2.3-fast/pro/image-to-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## MiniMax Hailuo 2.3 Fast
Source: https://raywake.si/explore/minimax/hailuo-2.3-fast/standard/image-to-video
Model ID: minimax/hailuo-2.3-fast/standard/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-2.3-fast/standard/image-to-video/api
Content updated: 2026-10-01

Describe how the subject and camera should move from the supplied first frame; leave the still to establish appearance.

Inputs and settings:
- image_url and prompt are required. duration offers 6 or 10 seconds; keep prompt within 2000 characters.
- Use prompt_optimizer to compare automatic prompt rewriting. This Raywake operation exposes neither an ending-frame input nor a resolution selector.

Review the result:
Check the opening-frame match, limb movement and background stability across the clip.

## MiniMax Hailuo 2.3 Pro
Source: https://raywake.si/explore/minimax/hailuo-2.3/pro/text-to-video
Model ID: minimax/hailuo-2.3/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-2.3/pro/text-to-video/api
Content updated: 2026-10-01

Plan a single action with a clear camera direction.

Inputs and settings:
- The default operation requires prompt. Describe the subject, action and surroundings; prompt_optimizer controls automatic rewriting of your brief.
- Choose image-to-video to animate an existing composition: it requires image_url and prompt. This Pro form exposes no duration or resolution selector.

Review the result:
Watch for changing faces, drifting objects and abrupt camera movement.

## MiniMax Hailuo 2.3 Standard
Source: https://raywake.si/explore/minimax/hailuo-2.3/standard/text-to-video
Model ID: minimax/hailuo-2.3/standard/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-2.3/standard/text-to-video/api
Content updated: 2026-10-01

Match the amount of movement to the selected clip length.

Inputs and settings:
- Text-to-video requires prompt; the alternate image-to-video operation also requires image_url. Describe motion within the source frame when using an image.
- duration offers 6 or 10 seconds. Use prompt_optimizer to control automatic prompt rewriting; compare the action before extending the clip.

Review the result:
Check that the action finishes naturally and the subject remains recognizable.

## MiniMax Music 3
Source: https://raywake.si/explore/minimax/music-3
Model ID: minimax/music-3
Provider documentation: https://fal.ai/models/minimax/music-3/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Write the musical direction and lyrics in the fields supplied by this version. Where audio settings are available, choose sample rate, bitrate and format.
- duration is an upper bound from 1–300 seconds; the model may stop earlier. Put each lyrics structure tag on its own line.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## MiniMax Speech 2.8 [HD]
Source: https://raywake.si/explore/minimax/speech-2.8-hd
Model ID: minimax/speech-2.8-hd
Provider documentation: https://fal.ai/models/fal-ai/minimax/speech-2.8-hd/api
Content updated: 2026-10-01

Write the actual script and use pauses deliberately.

Inputs and settings:
- This version takes prompt, with a 10000-character Raywake limit. voice_setting selects voice configuration; pronunciation_dict supports pronunciation adjustments.
- audio_setting, normalization_setting and voice_modify configure audio and voice treatment. The documented <#x#> pause notation belongs inside the spoken script.

Review the result:
Listen for pronunciation, pacing and whether expressive sounds fit the passage.

## MiniMax Speech-02 HD
Source: https://raywake.si/explore/minimax/speech-02-hd
Model ID: minimax/speech-02-hd
Provider documentation: https://fal.ai/models/fal-ai/minimax/speech-02-hd/api
Content updated: 2026-10-01

Prepare narration with readable punctuation and clearly written names.

Inputs and settings:
- Put the spoken script in text, not prompt. Raywake accepts up to 5000 characters; voice_setting and audio_setting hold voice and audio configuration.
- language_boost and pronunciation_dict can guide language handling and pronunciation. Review a short passage containing your difficult words before preparing the full narration.

Review the result:
Listen for mispronounced names, rushed pauses and unexpected number readings.

## MiniMax Speech-02 Turbo
Source: https://raywake.si/explore/minimax/speech-02-turbo
Model ID: minimax/speech-02-turbo
Provider documentation: https://fal.ai/models/fal-ai/minimax/speech-02-turbo/api
Content updated: 2026-10-03

Turn your text into speech with MiniMax Speech-02 Turbo and its voice controls.

Inputs and settings:
- Enter non-empty text. The credit hold covers the complete text, including multibyte characters; final billing uses recorded provider usage.
- Up to 5,000 characters. Set voice, emotion, speed, pitch, language boost, pronunciation dictionary and audio settings; output is saved as a URL.

Review the result:
Listen for names, pronunciation, pauses, clipping and the intended tone before publishing.

## MiniMax Voice Cloning
Source: https://raywake.si/explore/minimax/voice-clone
Model ID: minimax/voice-clone
Provider documentation: https://fal.ai/models/fal-ai/minimax/voice-clone/api
Content updated: 2026-10-01

Use a clear recording of the voice you have permission to reproduce.

Inputs and settings:
- audio_url is required. The official API asks for a sample of at least 10 seconds; optional text provides a preview script.
- model chooses the preview speech model. noise_reduction and need_volume_normalization treat the sample; accuracy is a text-validation threshold, not an overall quality score.

Review the result:
Compare the preview with the sample for pronunciation, tone and recording noise.

## MiniMax Hailuo 02 Fast
Source: https://raywake.si/explore/minimax/hailuo-02-fast/image-to-video
Model ID: minimax/hailuo-02-fast/image-to-video
Provider documentation: https://fal.ai/models/fal-ai/minimax/hailuo-02-fast/image-to-video/api
Content updated: 2026-10-01

Use a clear still and describe the movement that follows it, including the camera’s direction.

Inputs and settings:
- image_url and prompt are required. Choose duration as the string 6 or 10; prompt supports up to 2000 characters.
- prompt_optimizer can rewrite the instruction. Compare its effect using the same source; this Raywake route has no resolution or ending-frame control.

Review the result:
Compare the opening with the source; watch for warped faces, sliding objects and background drift.

## Minimax Music
Source: https://raywake.si/explore/minimax-music/v2
Model ID: minimax-music/v2
Provider documentation: https://fal.ai/models/fal-ai/minimax-music/v2/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Write the musical direction and lyrics in the fields supplied by this version. Where audio settings are available, choose sample rate, bitrate and format.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## Minimax Music 2.5
Source: https://raywake.si/explore/minimax-music/v2.5
Model ID: minimax-music/v2.5
Provider documentation: https://fal.ai/models/fal-ai/minimax-music/v2.5/api
Content updated: 2026-10-02

Write the words or musical direction and upload a reference if this operation requires one.

Inputs and settings:
- Write the musical direction and lyrics in the fields supplied by this version. Where audio settings are available, choose sample rate, bitrate and format.
- Provide lyrics, enable lyrics_optimizer, or choose is_instrumental for vocal-free music.

Review the result:
Listen to the complete audio and check pronunciation, lyrics, timing and sound quality.

## Minimax Music 2.6
Source: https://raywake.si/explore/minimax-music/v2.6
Model ID: minimax-music/v2.6
Provider documentation: https://fal.ai/models/fal-ai/minimax-music/v2.6/api
Content updated: 2026-10-01

Separate the musical style brief from lyrics; section tags help express verse and chorus structure.

Inputs and settings:
- prompt needs 10–2000 characters. Set is_instrumental for a vocal-free track; otherwise supply lyrics or use lyrics_optimizer with empty lyrics.
- Raywake accepts up to 3500 lyric characters and an audio_setting object. Keep song words in lyrics and instrumentation in prompt.

Review the result:
Listen for mispronounced words, uneven transitions and an abrupt ending; review the full arrangement.

## Moondream 2 Object Detection
Source: https://raywake.si/explore/moondream2/object-detection
Model ID: moondream2/object-detection
Provider documentation: https://fal.ai/models/fal-ai/moondream2/object-detection/api
Content updated: 2026-10-01

Locate a named object in a photo with Moondream 2 Object Detection. Use a specific object label rather than an instruction to edit the image.

Inputs and settings:
- Send image_url and the object label in object.
- Review the detected objects and the annotated image together. Keep the original dimensions when using coordinates downstream.

Review the result:
Check missed objects and incorrect detections manually; detection is not a segmentation cutout.

## MuseTalk
Source: https://raywake.si/explore/musetalk
Model ID: musetalk
Provider documentation: https://fal.ai/models/fal-ai/musetalk/api
Content updated: 2026-10-01

Choose footage with a visible mouth and prepare the audio you want it to follow.

Inputs and settings:
- Provide both source_video_url and audio_url. MuseTalk uses an existing video and audio track; this Raywake operation has no text prompt.
- Check that both media files open before submission. Listen while reviewing the full result, especially profile views, occlusions and pauses between spoken phrases.

Review the result:
Watch lips against consonants and check face edges during head movement.

## Nano Banana
Source: https://raywake.si/explore/nano-banana
Model ID: nano-banana
Provider documentation: https://fal.ai/models/fal-ai/nano-banana/api
Content updated: 2026-10-01

Specify composition, lighting and any exact text for a visual brief.

Inputs and settings:
- The default requires prompt. Editing additionally requires image_urls; explain what should change and what should stay.
- Choose aspect_ratio, output_format and num_images. Use the operation’s available ratios: auto appears on editing, but is absent from the default Raywake form.

Review the result:
Proofread lettering and compare preserved details with source images after editing.

## Nano Banana 2
Source: https://raywake.si/explore/nano-banana-2
Model ID: nano-banana-2
Provider documentation: https://fal.ai/models/fal-ai/nano-banana-2/api
Content updated: 2026-10-01

Set the composition before requesting a larger output.

Inputs and settings:
- The default starts with prompt. To alter existing images, select editing and supply image_urls; optional system_prompt carries broader style instructions.
- aspect_ratio includes auto and panoramic choices. resolution offers 0.5K, 1K, 2K or 4K; num_images and output_format control count and export format.

Review the result:
Inspect small lettering and object placement at the final output size.

## Nano Banana 2 Lite
Source: https://raywake.si/explore/google/nano-banana-2-lite
Model ID: google/nano-banana-2-lite
Provider documentation: https://fal.ai/models/google/nano-banana-2-lite/api
Content updated: 2026-10-01

Explore a composition before adapting it to a banner or portrait layout.

Inputs and settings:
- Send prompt and choose aspect_ratio; 4:1, 1:4, 8:1 and 1:8 are available. num_images allows 1–4 outputs.
- system_prompt can hold shared style instructions. To change an existing image, select the separate edit mode and supply image_urls.

Review the result:
Inspect lettering and subjects near the edges, especially with extreme aspect ratios.

## Nano Banana Pro
Source: https://raywake.si/explore/nano-banana-pro
Model ID: nano-banana-pro
Content updated: 2026-10-01

Develop a portrait or revise an existing image in the studio. Separate the parts you want to change from the parts you want to preserve; the editorial sample uses restrained styling and a clear focal point.

Suggested starting prompt (not the exact prompt behind the preview):
An editorial portrait in a black turtleneck, the face centered in the composition, natural skin texture, soft directional daylight and a quiet neutral backdrop.

Inputs and settings:
- Use a text prompt for a new image. For changes to an existing photo, select the editing mode and provide its image inputs; a prompt alone does not preserve a source image.
- Select aspect_ratio for your layout and resolution from the current schema. Start with one output using num_images, then request a quote before increasing resolution or the image count.
- Choose output_format for delivery. Save the seed and input settings when comparing variations; inspect generated lettering and portrait detail at full size.

Review the result:
Check facial details, hands and background edges. When editing, compare the result with the source image to confirm the elements you meant to preserve.

### Does Nano Banana Pro have a Raywake API?
Yes. Raywake exposes Nano Banana Pro with the text-to-image model ID nano-banana-pro. Its separate editing operation takes source images. Check the operation schema, request a quote, then start the job through POST /v1/generate.
Model schemas: https://raywake.si/docs/reference/models
### How much does Nano Banana Pro cost on Raywake?
Raywake charges prepaid credits per generation. Request a quote for the selected operation, resolution and image count; the studio and API share your balance. Prices and limits from a Google app or another provider do not describe your Raywake request.
Quotes and credits: https://raywake.si/docs/credits
### Where can I see actual Nano Banana Pro outputs?
Our dated perfume comparison shows one actual Nano Banana Pro output alongside FLUX.2 Pro, Seedream 4.5 and GPT Image 2.5 Sunburst. It records the common prompt, settings, dimensions, credits and elapsed time. One sample per model does not establish a quality winner.
See all four outputs: https://raywake.si/blog/one-api-image-video-audio#section-one-perfume-prompt-four-real-model-outputs

## Object Removal
Source: https://raywake.si/explore/object-removal
Model ID: object-removal
Provider documentation: https://fal.ai/models/fal-ai/object-removal/api
Content updated: 2026-10-01

Remove a named object using the default Object Removal operation. Use a precise description before switching to box or mask selection.

Inputs and settings:
- Send image_url and prompt for the default text-selection mode; other modes have separate inputs.
- model selects the quality option. mask_expansion extends the selection in pixels; compare nearby edges before increasing it.

Review the result:
Inspect replacement texture and shadows, particularly where the erased item overlapped another subject.

## Omni Zero
Source: https://raywake.si/explore/omni-zero
Model ID: omni-zero
Content updated: 2026-10-01

Use separate references to guide composition, style and identity. Choose references that describe compatible goals; the output is a new interpretation and may not preserve every detail from a source photograph.

Inputs and settings:
- Check image_url, composition_image_url, style_image_url and identity_image_url in the current schema. Each reference has a different purpose; a style reference does not replace an identity reference.
- Write a prompt that supports your references. Begin with one controlled change and keep the input images and selected settings when comparing results.

Review the result:
Compare the subject identity, pose, proportions and background against each reference. Inspect hands, faces and fine textures at the actual output size before selecting a result.

## Omnigen V2
Source: https://raywake.si/explore/omnigen-v2
Model ID: omnigen-v2
Provider documentation: https://fal.ai/models/fal-ai/omnigen-v2/api
Content updated: 2026-10-01

For multiple references, identify which image supplies each object or visual detail.

Inputs and settings:
- prompt is required; input_image_urls is optional for editing or combining references. Text-only requests can start without source images.
- text_guidance_scale and image_guidance_scale adjust text and reference influence separately. Keep cfg_range_start, cfg_range_end and scheduler unchanged while first refining the brief.

Review the result:
Check that referenced objects land in the intended scene and retain their defining features.

## Optimized Latent Consistency (SDv1.5)
Source: https://raywake.si/explore/lcm-sd15-i2i
Model ID: lcm-sd15-i2i
Provider documentation: https://fal.ai/models/fal-ai/lcm-sd15-i2i/api
Content updated: 2026-10-01

Explore a new palette or atmosphere while using your image as the starting point. The provider recommends 512×512 inputs.

Inputs and settings:
- Provide image_url and prompt. Adjust strength to vary the transformation; keep the same seed when comparing settings.
- Raywake exposes 1–12 inference steps and 1–8 images. Start with a small batch before changing guidance_scale.

Review the result:
Compare silhouettes and fine details with the original; a useful variation can still change the subject.

## Ovis Image
Source: https://raywake.si/explore/ovis-image
Model ID: ovis-image
Provider documentation: https://fal.ai/models/fal-ai/ovis-image/api
Content updated: 2026-10-01

Write the exact words and describe their placement relative to the subject.

Inputs and settings:
- prompt is required. Keep intended wording separate from the surrounding scene description; negative_prompt can describe unwanted details.
- Choose num_images and output_format. Adjust guidance_scale or num_inference_steps individually; keep acceleration and seed recorded so comparisons remain understandable.

Review the result:
Proofread every character and inspect spacing, hands and small shapes.

## P-Video 2 Pro
Source: https://raywake.si/explore/p-video-2-pro
Model ID: p-video-2-pro
Provider documentation: https://wiro.ai/models/pruna/p-video-2-pro
Content updated: 2026-10-01

Describe one action and its accompanying ambience.

Inputs and settings:
- prompt is required. inputImage and inputImageLast each accept one image; with a first frame, its shape determines framing rather than aspectRatio.
- Choose duration from 5, 10 or 15 seconds and resolution from 480p or 768p. mode and promptUpsampler select the generation recipe and prompt expansion.

Review the result:
Watch frame continuity and listen for sound that conflicts with the scene.

## Phota Enhance
Source: https://raywake.si/explore/phota/enhance
Model ID: phota/enhance
Provider documentation: https://fal.ai/models/fal-ai/phota/enhance/api
Content updated: 2026-10-01

Use an existing photograph when you want to refine its appearance.

Inputs and settings:
- image_url is required; this operation has no prompt. profile_ids optionally supplies existing profiles as identity-preservation candidates.
- Choose num_images and JPEG or PNG output_format. Start with one output and inspect skin, eyes and hair before requesting alternatives.

Review the result:
Compare the result with the original for facial changes and invented texture.

## Phota
Source: https://raywake.si/explore/phota
Model ID: phota
Provider documentation: https://fal.ai/models/fal-ai/phota/api
Content updated: 2026-10-01

Plan a portrait’s setting and framing before changing appearance.

Inputs and settings:
- Text generation requires prompt. Editing requires image_urls and prompt, with optional profile_ids; use only profile identifiers you already have.
- Choose aspect_ratio, num_images and output_format. resolution offers 1K or 4K, but official documentation limits 4K to the nb2 base_model.

Review the result:
Compare facial structure, hair and clothing with any supplied reference.

## Photo Restoration
Source: https://raywake.si/explore/image-apps-v2/photo-restoration
Model ID: image-apps-v2/photo-restoration
Provider documentation: https://fal.ai/models/fal-ai/image-apps-v2/photo-restoration/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## PhotoMaker
Source: https://raywake.si/explore/photomaker
Model ID: photomaker
Provider documentation: https://fal.ai/models/fal-ai/photomaker/api
Content updated: 2026-10-01

Choose clear references showing the same subject, then describe a new scene.

Inputs and settings:
- image_archive_url and prompt are required. Supply an image archive rather than a single-image URL; optional initial_image_url provides an additional starting composition.
- initial_image_strength controls noise applied to the starting image. style and style_strength guide treatment; change these separately from guidance_scale when comparing portraits.

Review the result:
Compare identity, facial proportions and unwanted changes to the source photographs.

## Pika 2.2
Source: https://raywake.si/explore/pika/v2.2/text-to-video
Model ID: pika/v2.2/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pika/v2.2/text-to-video/api
Content updated: 2026-10-01

Keep the action simple and assign each reference a clear role.

Inputs and settings:
- Default text-to-video needs prompt. Image-to-video adds image_url; Pikascenes requires image_urls and prompt to combine visual ingredients.
- Select resolution from 720p or 1080p. Text and Pikascenes expose aspect_ratio; negative_prompt names exclusions. The current Raywake forms do not expose duration.

Review the result:
Check subject continuity and whether combined objects stay distinct.

## PixVerse C1
Source: https://raywake.si/explore/pixverse/c1/text-to-video
Model ID: pixverse/c1/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/c1/text-to-video/api
Content updated: 2026-10-01

Describe the visible action and include sound cues when audio is enabled.

Inputs and settings:
- Default generation needs prompt. Image-to-video adds image_url; Transition requires first_image_url and end_image_url as well as prompt.
- duration supports 1–15 seconds; choose resolution and generate_audio_switch. aspect_ratio appears on text and transition operations; keep prompts concise because upstream limits UTF-8 bytes.

Review the result:
Watch continuity and listen for mismatched effects or unclear dialogue.

## PixVerse Extend
Source: https://raywake.si/explore/pixverse/extend
Model ID: pixverse/extend
Provider documentation: https://fal.ai/models/fal-ai/pixverse/extend/api
Content updated: 2026-10-01

Describe the next action using the position and direction visible in the final frames.

Inputs and settings:
- Provide video_url and a continuation prompt. Use resolution, style and negative_prompt to shape the added footage.
- The separate fast operation offers 360p, 540p or 720p; the default also offers 1080p. Keep seed and other settings when comparing attempts.

Review the result:
Watch the join for jumps in subject position, lighting and camera direction.

## PixVerse 3.5
Source: https://raywake.si/explore/pixverse/v3.5/text-to-video
Model ID: pixverse/v3.5/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v3.5/text-to-video/api
Content updated: 2026-10-01

Try one subject performing a clear action before adding a complicated camera move.

Inputs and settings:
- The default uses prompt. Select aspect_ratio and resolution; negative_prompt describes details to avoid.
- Image-to-video additionally needs image_url. Transition needs first_image_url and end_image_url, plus prompt; choose it when both endpoint frames matter.

Review the result:
Check whether the action follows your sequence and objects keep their shape.

## PixVerse 3.5 Fast
Source: https://raywake.si/explore/pixverse/v3.5/text-to-video/fast
Model ID: pixverse/v3.5/text-to-video/fast
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v3.5/text-to-video/fast/api
Content updated: 2026-10-01

Explore a short subject action with a fixed camera before testing another composition.

Inputs and settings:
- Start with prompt, then select aspect_ratio and resolution: 360p, 540p or 720p. style and negative_prompt are optional.
- For a supplied opening image, select the image-to-video fast operation and provide image_url plus prompt. Save seed when comparing settings.

Review the result:
Look for background drift and unintended changes in the subject during motion.

## PixVerse 4
Source: https://raywake.si/explore/pixverse/v4/text-to-video
Model ID: pixverse/v4/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v4/text-to-video/api
Content updated: 2026-10-01

Use a clear environment and one camera direction to explore an establishing shot.

Inputs and settings:
- Describe the scene in prompt. Set aspect_ratio and resolution; style, negative_prompt and seed are optional.
- Select image-to-video to start from an existing image. That operation requires image_url plus prompt and exposes camera_movement rather than aspect_ratio.

Review the result:
Check horizon stability, scale and whether the camera moves in the requested direction.

## PixVerse 4 Fast
Source: https://raywake.si/explore/pixverse/v4/text-to-video/fast
Model ID: pixverse/v4/text-to-video/fast
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v4/text-to-video/fast/api
Content updated: 2026-10-01

Try a landscape shot with one gentle camera movement.

Inputs and settings:
- Use prompt, aspect_ratio and resolution to define the default clip. Resolution choices are 360p, 540p and 720p.
- For an opening photograph, choose image-to-video fast and provide image_url plus prompt. camera_movement is available in that operation; style and seed are optional.

Review the result:
Look for warped buildings, an unstable horizon and motion that changes direction.

## PixVerse 4.5
Source: https://raywake.si/explore/pixverse/v4.5/text-to-video
Model ID: pixverse/v4.5/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v4.5/text-to-video/api
Content updated: 2026-10-01

Sketch a product reveal by separating the object action from the camera movement.

Inputs and settings:
- Use prompt for the default text-to-video operation. Select aspect_ratio, resolution and an optional style.
- Image-to-video requires image_url and offers camera_movement. Transition instead requires first_image_url and end_image_url; describe the change between them.

Review the result:
Inspect object edges and reflections as the camera changes position.

## PixVerse 4.5 Fast
Source: https://raywake.si/explore/pixverse/v4.5/text-to-video/fast
Model ID: pixverse/v4.5/text-to-video/fast
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v4.5/text-to-video/fast/api
Content updated: 2026-10-01

Test a reveal with one camera move and a simple background.

Inputs and settings:
- The default requires prompt and offers aspect_ratio plus 360p, 540p or 720p resolution. Use negative_prompt for unwanted details.
- The image-to-video fast operation requires image_url and prompt, and adds camera_movement. Keep the requested camera move consistent with the opening frame.

Review the result:
Watch for changes in perspective and accidental cuts during the reveal.

## PixVerse 5
Source: https://raywake.si/explore/pixverse/v5/text-to-video
Model ID: pixverse/v5/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v5/text-to-video/api
Content updated: 2026-10-01

Plan a single action with a clear beginning and end before adding several subjects.

Inputs and settings:
- Use the default with prompt, aspect_ratio and resolution. Optional style and negative_prompt guide appearance and exclusions.
- Image-to-video requires image_url plus prompt. Transition takes first_image_url and end_image_url plus prompt when you need defined opening and closing frames.

Review the result:
Check whether bodies, props and background details remain coherent across frames.

## PixVerse 5.5
Source: https://raywake.si/explore/pixverse/v5.5/text-to-video
Model ID: pixverse/v5.5/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v5.5/text-to-video/api
Content updated: 2026-10-01

Describe the action, camera and desired sound separately so the scene stays readable.

Inputs and settings:
- The default starts with prompt. Set aspect_ratio and resolution; generate_audio_switch requests audio when enabled.
- Image-to-video needs image_url. Transition needs first_image_url and end_image_url. Both also require prompt and offer the same audio switch.

Review the result:
Listen for misplaced effects and check whether motion matches the described action.

## PixVerse 5.6
Source: https://raywake.si/explore/pixverse/v5.6/text-to-video
Model ID: pixverse/v5.6/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v5.6/text-to-video/api
Content updated: 2026-10-01

Try an ambient scene and specify which sound belongs to which visible action.

Inputs and settings:
- Provide prompt, then choose aspect_ratio and resolution. generate_audio_switch controls the request for generated audio.
- Start from a photo using image_url in image-to-video, or connect two frames using first_image_url and end_image_url in transition. Keep seed when reviewing variations.

Review the result:
Play the full clip with sound and check for audio events without a matching action.

## PixVerse 6
Source: https://raywake.si/explore/pixverse/v6/text-to-video
Model ID: pixverse/v6/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/pixverse/v6/text-to-video/api
Content updated: 2026-10-01

Match the requested action to the chosen duration; keep a short clip focused on one event.

Inputs and settings:
- The default uses prompt, aspect_ratio and duration from 1 to 15 seconds. Select resolution and generate_audio_switch as needed.
- Image-to-video needs image_url, transition needs first_image_url and end_image_url, and extend needs video_url. Each also requires prompt and offers duration.

Review the result:
Review the complete action, ending and sound timing rather than judging only the opening frame.

## Pixelcut Background Remover
Source: https://raywake.si/explore/pixelcut/background-removal
Model ID: pixelcut/background-removal
Provider documentation: https://fal.ai/models/pixelcut/background-removal/api
Content updated: 2026-10-01

Prepare an image cutout with Pixelcut Background Remover. Start from a clear source whose foreground you can inspect around fine edges.

Inputs and settings:
- Provide image_url with a JPEG or PNG source. This Raywake operation exposes the source image as its user input.
- Inspect the delivered cutout rather than relying on the catalog preview. Keep transparency when exporting for a composite; JPEG cannot retain it.

Review the result:
Check jewelry gaps, hair and translucent areas against the original. Test the result on the background where it will be used.

## Playground v2.5
Source: https://raywake.si/explore/playground-v25
Model ID: playground-v25
Provider documentation: https://fal.ai/models/fal-ai/playground-v25/api
Content updated: 2026-10-01

Explore composition and lighting before refining smaller objects.

Inputs and settings:
- The default requires prompt. num_images allows 1–8; num_inference_steps allows 1–50. Change guidance_scale gradually while keeping seed fixed.
- Image-to-image needs image_url and prompt; inpainting also needs mask_url. These operations offer strength and their own step limits. Choose format as jpeg or png.

Review the result:
Inspect object relationships and check that requested details are not replaced by unrelated decoration.

## Pony V7
Source: https://raywake.si/explore/pony-v7
Model ID: pony-v7
Provider documentation: https://fal.ai/models/fal-ai/pony-v7/api
Content updated: 2026-10-01

Give the subject, pose, material and lighting clear roles in the prompt.

Inputs and settings:
- Start with prompt and num_images of 1 or 2. num_inference_steps accepts 20–50; guidance_scale accepts 0–20.
- Use output_format for jpeg or png. Keep seed and noise_source unchanged while adjusting one setting so the comparison stays meaningful.

Review the result:
Inspect anatomy, repeated details and whether the requested pose is recognizable.

## Post Processing
Source: https://raywake.si/explore/post-processing/blur
Model ID: post-processing/blur
Content updated: 2026-10-01

Apply a blur effect to an existing image when you want a softer or more painterly result. This is an image-processing step, not a prompt-based scene generator or a reliable privacy redaction tool.

Inputs and settings:
- Provide the source using image_url and select the blur method from the current schema. Gaussian and Kuwahara produce different visual effects.
- Begin with a modest blur setting. Adjust the supported radius and sigma controls one at a time and retain an untouched source image.

Review the result:
Compare edges, text and facial features with the original. Blurring can leave identifying details visible; inspect the exported result at its intended display size.

## Post Processing Sharpen
Source: https://raywake.si/explore/post-processing/sharpen
Model ID: post-processing/sharpen
Provider documentation: https://fal.ai/models/fal-ai/post-processing/sharpen/api
Content updated: 2026-10-03

Sharpen an image with basic, smart or contrast-adaptive sharpening.

Inputs and settings:
- Upload image_url and choose sharpen_mode. Set its strength/radius controls; compare at full size to avoid halos.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Qwen Image 3 Image Editing
Source: https://raywake.si/explore/alibaba/qwen-image-3/edit
Model ID: alibaba/qwen-image-3/edit
Provider documentation: https://fal.ai/models/alibaba/qwen-image-3/edit/api
Content updated: 2026-10-01

Change a background, object or clothing while specifying which source details should remain.

Inputs and settings:
- Provide image_urls and prompt. The provider accepts 1–3 references; refer to their order when combining details.
- Keep prompt within 5000 characters and negative_prompt within 500. Select 1–6 outputs with num_images and JPEG, PNG or WebP with output_format.

Review the result:
Compare faces, object geometry and untouched areas against the references at full size.

## Qwen Image 3.0
Source: https://raywake.si/explore/qwen-image-3-0
Model ID: qwen-image-3-0
Provider documentation: https://wiro.ai/models/alibaba/qwen-image-3-0
Content updated: 2026-10-01

Describe layout, subject and any requested wording in a clear order.

Inputs and settings:
- Provide prompt of up to 4,500 characters. This operation accepts n=1; select resolution as 1K or 2K.
- Optional ratio and negativePrompt refine the request. Keep seed when comparing attempts; describe all content in text for this operation.

Review the result:
Read every letter and check alignment, spacing and the placement of objects.

## Qwen Add Background
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/add-background
Model ID: qwen-image-edit-plus-lora-gallery/add-background
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/add-background/api
Content updated: 2026-10-01

Place a clean-background subject in a new setting with Qwen Add Background. Start with an object that is easy to separate from its surroundings.

Inputs and settings:
- Supply image_urls with a white or clean-background source. Describe the new environment in prompt.
- Adjust lora_scale for the effect and save seed when comparing backgrounds.

Review the result:
Check subject edges, contact shadows and scale against the new background.

## Qwen Face to Full Portrait
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/face-to-full-portrait
Model ID: qwen-image-edit-plus-lora-gallery/face-to-full-portrait
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/face-to-full-portrait/api
Content updated: 2026-10-01

Expand a close-up face into a portrait concept with Qwen Face to Full Portrait. Specify the clothing, pose and setting you want.

Inputs and settings:
- Put the face reference in image_urls and describe the intended portrait in prompt.
- Keep the reference and seed unchanged while comparing lora_scale settings.

Review the result:
Compare facial likeness, neck and clothing boundaries with the reference; the added body and scene are generated.

## Qwen Group Photo
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/group-photo
Model ID: qwen-image-edit-plus-lora-gallery/group-photo
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/group-photo/api
Content updated: 2026-10-01

Combine individual portraits into a shared scene with Qwen Group Photo. Choose references where each face is clearly visible.

Inputs and settings:
- Supply two or more portrait references in image_urls. Describe the setting and arrangement in prompt.
- Start with one output using num_images. Save reference order and seed for comparisons.

Review the result:
Check every face, subject count and relative height. The composite does not document a real gathering.

## Qwen Integrate Product
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/integrate-product
Model ID: qwen-image-edit-plus-lora-gallery/integrate-product
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/integrate-product/api
Content updated: 2026-10-01

Blend a product into its surrounding scene with Qwen Integrate Product. Use a source that shows the object and the intended environment.

Inputs and settings:
- Send image_urls and use prompt to describe the intended integration.
- Change lora_scale separately from guidance_scale when comparing how the product blends into the scene.

Review the result:
Inspect perspective, contact shadows, packaging shape and any brand lettering.

## Qwen Lighting Restoration
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/lighting-restoration
Model ID: qwen-image-edit-plus-lora-gallery/lighting-restoration
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/lighting-restoration/api
Content updated: 2026-10-01

Rework image lighting with Qwen Lighting Restoration. Compare the result with the source to judge the tonal changes.

Inputs and settings:
- Send the source through image_urls. This route has no prompt or lora_scale field.
- Save seed and adjust guidance_scale independently of num_inference_steps.

Review the result:
Check highlights, shadow transitions and color casts; generated restoration cannot prove the original lighting.

## Qwen Multiple Angles
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/multiple-angles
Model ID: qwen-image-edit-plus-lora-gallery/multiple-angles
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/multiple-angles/api
Content updated: 2026-10-01

Explore a different viewpoint of a source scene with Qwen Multiple Angles. Change one camera control at a time.

Inputs and settings:
- Send image_urls. Positive rotate_right_left turns left; negative values turn right.
- Use move_forward for closer framing and vertical_angle for elevation. wide_angle_lens adds the documented lens effect.

Review the result:
Inspect newly exposed surfaces and object geometry; this is a generated viewpoint, not a calibrated 3D reconstruction.

## Qwen Next Scene
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/next-scene
Model ID: qwen-image-edit-plus-lora-gallery/next-scene
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/next-scene/api
Content updated: 2026-10-01

Plan a following storyboard frame with Qwen Next Scene. Describe the transition from the supplied image to another still composition.

Inputs and settings:
- Send image_urls and begin prompt with the documented Next Scene: prefix.
- Describe a camera or framing change. This image route returns a still, not a video clip.

Review the result:
Compare character continuity, scene layout and newly revealed details with the preceding frame.

## Qwen Remove Element
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/remove-element
Model ID: qwen-image-edit-plus-lora-gallery/remove-element
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/remove-element/api
Content updated: 2026-10-01

Erase a described scene element with Qwen Remove Element. Specify which item should disappear when several similar objects are present.

Inputs and settings:
- Send image_urls and name the element in prompt.
- Save seed while comparing lora_scale. Keep the instruction focused on the removal.

Review the result:
Check the filled area and nearby subjects for unwanted replacements or leftover shadows.

## Qwen Remove Lighting
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/remove-lighting
Model ID: qwen-image-edit-plus-lora-gallery/remove-lighting
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/remove-lighting/api
Content updated: 2026-10-01

Neutralize lighting effects in an image with Qwen Remove Lighting. Use the result as a visual editing reference.

Inputs and settings:
- Supply image_urls; this operation has no text prompt or lora_scale control.
- Keep seed fixed when comparing guidance_scale and inference-step changes.

Review the result:
Inspect residual shadows and altered material colors. Do not treat the generated image as measured surface reflectance.

## Qwen Shirt Design
Source: https://raywake.si/explore/qwen-image-edit-plus-lora-gallery/shirt-design
Model ID: qwen-image-edit-plus-lora-gallery/shirt-design
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-edit-plus-lora-gallery/shirt-design/api
Content updated: 2026-10-01

Mock up a graphic on a worn shirt with Qwen Shirt Design. Keep the garment photo and design reference separate.

Inputs and settings:
- In image_urls, place the person wearing the shirt first and the design or logo second.
- Use prompt to describe placement and lora_scale to adjust the effect.

Review the result:
Check lettering, design proportions and folds. A mockup is not a production-ready print file.

## Qwen Image Max
Source: https://raywake.si/explore/qwen-image-max/text-to-image
Model ID: qwen-image-max/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/qwen-image-max/text-to-image/api
Content updated: 2026-10-01

Specify the subject and layout, and write any desired image text explicitly.

Inputs and settings:
- The default requires prompt, up to 800 characters. Use num_images from 1–4 and output_format as jpeg, png or webp.
- For edits, choose the edit operation with image_urls of up to three references and prompt. It permits num_images from 1–6. Both offer enable_prompt_expansion, negative_prompt and seed.

Review the result:
Read the generated wording and check object counts and composition.

## Qwen Image Max Uncensored
Source: https://raywake.si/explore/qwen-image-max-uncensored
Model ID: qwen-image-max-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/qwen-image-max-uncensored
Content updated: 2026-10-01

When using references, identify Image 1, Image 2 and Image 3 in the edit instructions.

Inputs and settings:
- Provide prompt of up to 800 characters; optional inputImage accepts up to three references. negativePrompt accepts up to 500 characters.
- Text-only runs return one image. samples and resolution apply when a reference is supplied; resolution choices are 1k and 2k. Keep seed for comparisons.

Review the result:
Compare faces, object positions and requested changes with each reference.

## Realistic Vision
Source: https://raywake.si/explore/realistic-vision
Model ID: realistic-vision
Provider documentation: https://fal.ai/models/fal-ai/realistic-vision/api
Content updated: 2026-10-01

Describe a specific lighting setup, camera viewpoint and subject rather than relying on a quality label.

Inputs and settings:
- Provide prompt and optional negative_prompt. Start with num_images=1; guidance_scale is 0–20 and num_inference_steps is 1–70.
- format selects jpeg or png. Keep seed when comparing lighting changes; expand_prompt can add detail, while loras and embeddings are optional advanced inputs.

Review the result:
Inspect skin, hands, reflections and contact shadows at full size.

## Recraft Vectorize
Source: https://raywake.si/explore/recraft/vectorize
Model ID: recraft/vectorize
Provider documentation: https://fal.ai/models/fal-ai/recraft/vectorize/api
Content updated: 2026-10-03

Convert a PNG, JPG or WebP image into a scalable SVG file.

Inputs and settings:
- Upload a file smaller than 5 MB, below 16 MP, with each side from 257 to 4095 pixels.
- Preview the result, then download the SVG to continue editing in your design tool.

Review the result:
Inspect path edges, small details and text before using the SVG in production.

## Recraft Upscale
Source: https://raywake.si/explore/recraft/upscale/crisp
Model ID: recraft/upscale/crisp
Provider documentation: https://fal.ai/models/fal-ai/recraft/upscale/crisp/api
Content updated: 2026-10-01

Enlarge an existing image with the Crisp upscaler. Start with a clean source; this operation does not accept a written prompt to replace the scene.

Inputs and settings:
- Provide image_url. The provider documentation specifies PNG input, so prepare a PNG source rather than assuming every format is accepted.

Review the result:
Inspect straight edges, lettering and small textures at full size. Compare the source and output at the same displayed size so enlargement alone does not look like a quality gain.

## Recraft V3
Source: https://raywake.si/explore/recraft/v3/text-to-image
Model ID: recraft/v3/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/recraft/v3/text-to-image/api
Content updated: 2026-10-01

Describe the main shape, its placement and the intended visual style.

Inputs and settings:
- Provide prompt of up to 1,000 characters. Optional style chooses the visual treatment; colors supplies preferred colors.
- Keep the composition description stable while changing style or palette. Use enable_safety_checker as provided by the form; this operation starts from text.

Review the result:
Review text, silhouettes and color use; inspect the returned file before treating it as a finished design.

## Recraft V4
Source: https://raywake.si/explore/recraft/v4/text-to-image
Model ID: recraft/v4/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/recraft/v4/text-to-image/api
Content updated: 2026-10-01

Develop an illustration or product scene from a written brief with preferred colors. Define the subject and layout first, then use the palette to guide the visual direction.

Inputs and settings:
- Provide prompt. colors accepts RGB color objects; background_color describes the preferred background separately.
- Review nested color fields in the complete schema. Keep the brief and palette with your selected output so you can compare another version.

Review the result:
Check actual colors, small lettering and empty space before placing the image in a design. Preferred colors are guidance, not a guarantee of exact brand-color reproduction.

## Recraft V4 Pro
Source: https://raywake.si/explore/recraft/v4/pro/text-to-image
Model ID: recraft/v4/pro/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/recraft/v4/pro/text-to-image/api
Content updated: 2026-10-01

Build a composition with Recraft V4 Pro from a prompt and preferred palette. Keep your subject and color choices fixed when testing it alongside the standard V4 endpoint.

Inputs and settings:
- Send prompt and optionally colors and background_color as RGB objects. Check their nested fields in the full schema.
- Quote Pro separately and record the selected layout. Compare the actual output, price and turnaround time instead of assuming an advantage from the Pro label.

Review the result:
Inspect textures, silhouettes and requested lettering. Compare both outputs at a common display size; do not mistake enlargement for improved detail.

## Recraft V4 Custom Style
Source: https://raywake.si/explore/recraft/v4/style/text-to-image
Model ID: recraft/v4/style/text-to-image
Provider documentation: https://fal.ai/models/recraft/v4/style/text-to-image/api
Content updated: 2026-10-03

Generate images in a reusable custom style with Recraft V4 Standard or Pro.

Inputs and settings:
- Supply a UUID style_id or one to ten reference image_urls (PNG, JPG or WebP). New style creation adds $0.005 to provider cost; existing style IDs avoid that addition.
- Choose output size, palette and background color. style_match can be precise or flexible. Read the full quote before starting.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Recraft V4.1 Flash
Source: https://raywake.si/explore/recraft/v4.1/flash/text-to-image
Model ID: recraft/v4.1/flash/text-to-image
Provider documentation: https://fal.ai/models/recraft/v4.1/flash/text-to-image/api
Content updated: 2026-10-01

Explore a visual concept by stating its subject, arrangement and dominant colors.

Inputs and settings:
- Provide prompt of up to 10,000 characters. colors sets preferred colors; background_color specifies a preferred background.
- Adjust the palette or background while keeping the scene description steady. This Raywake operation offers enable_safety_checker alongside those controls.

Review the result:
Check color placement, small lettering and edges before using the image in a layout.

## Recraft V4.1
Source: https://raywake.si/explore/recraft/v4.1/text-to-image
Model ID: recraft/v4.1/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/recraft/v4.1/text-to-image/api
Content updated: 2026-10-01

Specify foreground and background relationships, materials and the intended palette.

Inputs and settings:
- The default requires prompt, with a 10,000-character limit. colors and background_color are optional palette controls.
- Utility is a separate text-to-image operation with the same fields on Raywake. Compare the selected operation and settings before running; enable_safety_checker is also available.

Review the result:
Check whether focal details remain distinct and the chosen colors support the composition.

## Recraft V4.1 Pro
Source: https://raywake.si/explore/recraft/v4.1/pro/text-to-image
Model ID: recraft/v4.1/pro/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/recraft/v4.1/pro/text-to-image/api
Content updated: 2026-10-01

Describe the hierarchy of a campaign image, including the focal object and surrounding space.

Inputs and settings:
- Start with prompt of up to 10,000 characters. colors and background_color express palette preferences; enable_safety_checker is optional.
- The separate Utility Pro operation also starts from text and exposes the same four fields here. Choose the operation first, then compare the quoted settings.

Review the result:
Inspect fine edges and lettering, and verify actual file dimensions before arranging the final layout.

## Recraft V4.1 Vector
Source: https://raywake.si/explore/recraft/v4.1/text-to-vector
Model ID: recraft/v4.1/text-to-vector
Provider documentation: https://fal.ai/models/fal-ai/recraft/v4.1/text-to-vector/api
Content updated: 2026-10-03

Create SVG illustrations and icons from a text prompt with optional color controls.

Inputs and settings:
- Describe a single icon or illustration, choose the image size, and optionally set RGB colors and background color.
- Each request produces one SVG. Download the vector file to resize it or edit its paths in your design tool.

Review the result:
Inspect path edges, small details and text before using the SVG in production.

## Rembg Enhance (Remove Background Enhance)
Source: https://raywake.si/explore/smoretalk-ai/rembg-enhance
Model ID: smoretalk-ai/rembg-enhance
Provider documentation: https://fal.ai/models/smoretalk-ai/rembg-enhance/api
Content updated: 2026-10-01

Remove a source image’s background with Rembg Enhance. This route takes a single image input for segmentation.

Inputs and settings:
- Supply image_url; the current form has no prompt or extra tuning controls.
- Review your returned cutout rather than the catalog preview; preserve transparency when saving for compositing.

Review the result:
Compare thin details and foreground color near edges with the original.

## Remove Background
Source: https://raywake.si/explore/imageutils/rembg
Model ID: imageutils/rembg
Provider documentation: https://fal.ai/models/fal-ai/imageutils/rembg/api
Content updated: 2026-10-01

Create a subject cutout with the Remove Background utility. Start with a photo whose foreground is clearly visible.

Inputs and settings:
- Send image_url; there is no selection prompt.
- Use crop_to_bbox if the output should be cropped around the subject. Keep it off when the original canvas matters.

Review the result:
Inspect fine edges and transparent gaps against a contrasting background; preserve transparency when exporting.

## Runway Gen-4 Turbo
Source: https://raywake.si/explore/runway-gen4-turbo
Model ID: runway-gen4-turbo
Provider documentation: https://www.eachlabs.ai/runway/gen4/gen4-turbo
Content updated: 2026-10-01

Add a camera move or a simple action to an established composition.

Inputs and settings:
- Provide image and prompt. Describe what moves and how the camera follows; the image establishes the scene.
- Select duration of 5 or 10 seconds and a supported aspect_ratio. Record seed when comparing versions; Raywake’s current controls do not include an audio switch.

Review the result:
Watch for identity changes, unstable edges and movement that contradicts the photo.

## SDXL ControlNet Union
Source: https://raywake.si/explore/sdxl-controlnet-union
Model ID: sdxl-controlnet-union
Provider documentation: https://fal.ai/models/fal-ai/sdxl-controlnet-union/api
Content updated: 2026-10-01

Combine structural controls for an SDXL scene with ControlNet Union. Begin with one control type so its effect is easy to judge.

Inputs and settings:
- Send prompt and the matching control URL, such as depth_image_url or canny_image_url.
- Use the corresponding preprocess switch for that input. Different editing modes require their own image or mask fields.

Review the result:
Check alignment between control maps and whether added controls conflict with one another.

## SadTalker
Source: https://raywake.si/explore/sadtalker
Model ID: sadtalker
Provider documentation: https://fal.ai/models/fal-ai/sadtalker/api
Content updated: 2026-10-01

Use clear speech and a face with unobstructed mouth and eyes.

Inputs and settings:
- Default mode requires source_image_url and driven_audio_url. The reference operation additionally requires reference_pose_video_url to supply pose motion.
- expression_scale adjusts expression; still_mode reduces head movement with full preprocessing. Change preprocess and face_model_resolution separately so framing and facial detail remain easy to judge.

Review the result:
Check lip timing, facial expression and head movement against the recording.

## SAM 3.1 Image
Source: https://raywake.si/explore/sam-3-1/image
Model ID: sam-3-1/image
Provider documentation: https://fal.ai/models/fal-ai/sam-3-1/image/api
Content updated: 2026-10-01

Prepare an object selection for a later cutout or editing step.

Inputs and settings:
- Supply image_url, then identify the target with prompt, point_prompts or box_prompts. Boxes use x_min, y_min, x_max and y_max coordinates.
- apply_mask controls the masked preview. return_multiple_masks, include_scores and include_boxes expose additional results; choose output_format for delivery. This operation selects regions rather than generating a new scene.

Review the result:
Inspect thin edges, holes and nearby objects included in the mask.

## Sana
Source: https://raywake.si/explore/sana
Model ID: sana
Provider documentation: https://fal.ai/models/fal-ai/sana/api
Content updated: 2026-10-01

Build an illustration or visual concept with a clear subject and style.

Inputs and settings:
- Write prompt and use negative_prompt for specific unwanted elements. Begin with one output; num_images supports 1–4.
- The provider starts at 18 steps and guidance_scale 5. Adjust these separately, record seed, and choose JPEG or PNG with output_format. style_name controls the visual style.

Review the result:
Check the main subject, repeated details and requested text at full size.

## Sana Sprint
Source: https://raywake.si/explore/sana/sprint
Model ID: sana/sprint
Provider documentation: https://fal.ai/models/fal-ai/sana/sprint/api
Content updated: 2026-10-01

Compare alternative scene briefs before refining the chosen direction.

Inputs and settings:
- Start with a focused prompt; negative_prompt can exclude unwanted elements. Keep num_images at one while comparing changes.
- Sprint’s provider example uses two inference steps; Raywake accepts 1–20. Keep seed fixed and adjust guidance_scale separately. style_name and output_format control style and JPEG/PNG delivery.

Review the result:
Check object shapes, texture and whether the composition follows your brief.

## Sana v1.5 1.6B
Source: https://raywake.si/explore/sana/v1.5/1.6b
Model ID: sana/v1.5/1.6b
Provider documentation: https://fal.ai/models/fal-ai/sana/v1.5/1.6b/api
Content updated: 2026-10-01

Explore a scene’s subject, lighting and style using a written brief.

Inputs and settings:
- Describe the scene in prompt and exclusions in negative_prompt. style_name lets you specify a visual treatment; this operation does not edit a source photo.
- Use num_images for 1–4 outputs. The API starts at 18 steps and guidance_scale 5; record seed and change one setting at a time. Select JPEG or PNG with output_format.

Review the result:
Review small objects and background details before selecting a result.

## Sana v1.5 4.8B
Source: https://raywake.si/explore/sana/v1.5/4.8b
Model ID: sana/v1.5/4.8b
Provider documentation: https://fal.ai/models/fal-ai/sana/v1.5/4.8b/api
Content updated: 2026-10-01

Test a detailed material or environment description with a defined focal point.

Inputs and settings:
- Use prompt for subject and setting, then negative_prompt for exclusions. Choose style_name deliberately rather than mixing several conflicting styles.
- The API starts at 18 steps and guidance_scale 5. Raywake allows 1–50 steps and 1–4 images; compare settings with a fixed seed and export JPEG or PNG through output_format.

Review the result:
Inspect perspective, fine textures and secondary objects in the result.

## Seedance 1 Pro
Source: https://raywake.si/explore/bytedance/seedance/v1/pro/text-to-video
Model ID: bytedance/seedance/v1/pro/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/bytedance/seedance/v1/pro/text-to-video/api
Content updated: 2026-10-01

Build a short action with intentional framing and camera movement.

Inputs and settings:
- Text mode requires prompt. Image mode adds image_url and optionally end_image_url for a target final frame.
- Choose 2–12 seconds and 480p/720p/1080p. Use camera_fixed when appropriate; num_frames overrides duration, so avoid conflicting timing instructions.

Review the result:
Check the full clip for changing objects, inconsistent faces and abrupt motion.

## Seedance 2 Fast
Source: https://raywake.si/explore/bytedance/seedance-2.0/fast/text-to-video
Model ID: bytedance/seedance-2.0/fast/text-to-video
Provider documentation: https://fal.ai/models/bytedance/seedance-2.0/fast/text-to-video/api
Content updated: 2026-10-01

Explore action and camera ideas before assembling a longer edit.

Inputs and settings:
- Text mode requires prompt. Image mode adds image_url and optional end_image_url; reference mode accepts up to nine image_urls.
- Choose 4–15 seconds and 480p or 720p. Set aspect_ratio and generate_audio; inspect the chosen codec and bitrate_mode in the complete form.

Review the result:
Review shot transitions, object continuity and any dialogue timing throughout the clip.

## Seedance 2 Mini
Source: https://raywake.si/explore/bytedance/seedance-2.0/mini/text-to-video
Model ID: bytedance/seedance-2.0/mini/text-to-video
Provider documentation: https://fal.ai/models/bytedance/seedance-2.0/mini/text-to-video/api
Content updated: 2026-10-01

Try compact scene ideas with explicit action, camera direction and sound cues.

Inputs and settings:
- Text mode uses prompt. Image mode requires image_url too; reference mode accepts optional image_urls, up to nine.
- Choose 4–15 seconds, 480p or 720p and aspect_ratio. Set generate_audio and review codec; Mini does not expose bitrate_mode.

Review the result:
Watch for changes in subject identity, abrupt transitions and mismatched sound.

## Seedance 2
Source: https://raywake.si/explore/bytedance/seedance-2.0/text-to-video
Model ID: bytedance/seedance-2.0/text-to-video
Provider documentation: https://fal.ai/models/bytedance/seedance-2.0/text-to-video/api
Content updated: 2026-10-01

Describe action and camera direction for a scene, adding speech or ambient sound when needed.

Inputs and settings:
- Text mode requires prompt. Image mode also needs image_url and can use end_image_url; reference mode accepts up to nine optional image_urls.
- Choose 4–15 seconds and 480p/720p/1080p. Set aspect_ratio and generate_audio; use the current codec and bitrate_mode choices for export.

Review the result:
Review character continuity, the order of actions and speech synchronization across the complete clip.

## Seedance 2 US
Source: https://raywake.si/explore/bytedance/seedance-2.0/us/text-to-video
Model ID: bytedance/seedance-2.0/us/text-to-video
Provider documentation: https://fal.ai/models/bytedance/seedance-2.0/us/text-to-video/api
Content updated: 2026-10-01

Plan one continuous shot. Use a starting image when its framing must carry into the video, or reference images when subject appearance matters.

Inputs and settings:
- Choose text, image or reference mode. Text mode needs prompt; image mode also needs image_url and can take end_image_url. Reference mode accepts up to nine image_urls.
- Raywake currently supports explicit durations of 4–15 seconds at 480p or 720p. Select an aspect ratio for text/reference mode; set generate_audio deliberately.
- Video/audio references and 1080p/4K are unavailable in this release. Encoding uses standard bitrate and automatic codec. Request a quote for the complete input before generating.

Review the result:
Watch the full clip for changing faces, objects and background details. Check generated speech and effects before sharing.

## Seedance 2.5
Source: https://raywake.si/explore/seedance-2-5
Model ID: bytedance/seedance-2.5/text-to-video
Content updated: 2026-10-01

Build a short atmospheric video around a simple action. The jellyfish example uses one subject and a restrained background to make movement easy to read.

Suggested starting prompt (not the exact prompt behind the preview):
A luminous jellyfish drifts slowly through deep blue water. Its tentacles move gently as the camera follows at a calm, constant distance.

Inputs and settings:
- The featured text-to-video endpoint starts with a text prompt. Use an image-to-video endpoint when the opening frame must come from your photo.
- Describe one action and camera movement, then choose duration, resolution and aspect_ratio from the current studio options. A longer clip needs a movement that can remain coherent throughout.
- Decide whether to include sound using generate_audio where available. Review the clip with sound enabled, and quote the selected settings before submitting.

Review the result:
Look for smooth movement, consistent shapes and a useful opening and ending frame. Trim the result in an editor if you need a seamless loop.

### How should I compare Seedance 2.5 costs?
Request quotes for the same task, duration, resolution and audio setting when comparing available operations. Raywake uses credits per generation; a subscription to another service is a separate offer. Review your actual output against the brief before deciding whether a run was worth its cost.
Understand generation charges: https://raywake.si/docs/credits

## Seedance 2.5 US
Source: https://raywake.si/explore/bytedance/seedance-2.5/us/text-to-video
Model ID: bytedance/seedance-2.5/us/text-to-video
Provider documentation: https://fal.ai/models/bytedance/seedance-2.5/us/text-to-video/api
Content updated: 2026-10-01

Describe an action that can continue through the selected duration. Use a starting photo for fixed framing and image references for a subject-led scene.

Inputs and settings:
- Select text, image or reference mode. Image mode requires image_url, optionally takes end_image_url and follows the input aspect ratio automatically. Reference mode uses image_urls.
- Choose 4–30 seconds and 480p, 720p or 1080p. Choose an explicit aspect ratio for text/reference mode and select generate_audio. Automatic duration is unavailable.
- Reference mode currently supports image guidance; editing, extension and video/audio references are unavailable. Encoding is fixed to standard bitrate and automatic codec. 1080p has a higher token rate; review the quote before submission.

Review the result:
Check subject consistency and camera movement from start to finish. Review the audio separately and confirm that the output fits your intended placement.

## Seedream 5 Flash
Source: https://raywake.si/explore/bytedance/seedream/v5/flash/text-to-image
Model ID: bytedance/seedream/v5/flash/text-to-image
Provider documentation: https://fal.ai/models/bytedance/seedream/v5/flash/text-to-image/api
Content updated: 2026-10-01

Try visual directions for an illustration or product composition, then refine selected source images.

Inputs and settings:
- Text mode requires prompt; editing additionally requires image_urls. Describe concrete changes rather than a general request to improve quality.
- Select 1–6 results with num_images and JPEG or PNG with output_format. Keep the brief focused when comparing versions; no seed control is exposed.

Review the result:
Inspect small lettering, object boundaries and changes to details you asked to preserve.

## Seedream 5 Lite
Source: https://raywake.si/explore/bytedance/seedream/v5/lite/text-to-image
Model ID: bytedance/seedream/v5/lite/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/bytedance/seedream/v5/lite/text-to-image/api
Content updated: 2026-10-01

Develop a visual concept and describe a targeted revision when the source composition matters.

Inputs and settings:
- Text mode needs prompt. Editing needs prompt and image_urls; refer clearly to the relevant source image.
- Use num_images for 1–6 results and start with one. Raywake currently exposes no size, format or seed controls for these operations.

Review the result:
Check requested wording, repeated details and whether editing preserved the intended objects.

## Seedream 4.5 Uncensored
Source: https://raywake.si/explore/seedream-v4-5-uncensored
Model ID: seedream-v4-5-uncensored
Provider documentation: https://wiro.ai/models/bytedance/seedream-v4-5-uncensored
Content updated: 2026-10-01

Use text alone for a new scene; add references for targeted image changes.

Inputs and settings:
- prompt is required. inputImage accepts up to 14 references; state what to preserve and what to change when editing.
- Raywake offers resolution 2k or 4k and supported aspectRatio values. maxImages is fixed at one here, despite broader upstream examples; review watermark before submitting.

Review the result:
Compare retained objects, faces and lettering with the supplied references.

## Seedream 5 Lite Uncensored
Source: https://raywake.si/explore/seedream-v5-lite-uncensored
Model ID: seedream-v5-lite-uncensored
Provider documentation: https://wiro.ai/models/bytedance/seedream-v5-lite-uncensored
Content updated: 2026-10-01

Combine a specific instruction with references when the existing subject matters.

Inputs and settings:
- Use prompt alone for generation, or add an inputImage array for editing and composition. Up to 14 reference images are accepted.
- Choose resolution 2k or 3k and a supported aspectRatio. Raywake limits maxImages to one; retain that limit even when the provider’s examples describe batches. Check watermark separately.

Review the result:
Check reference identity, unwanted changes and consistency between combined objects.

## Seedream 5 Pro
Source: https://raywake.si/explore/bytedance/seedream/v5/pro/text-to-image
Model ID: bytedance/seedream/v5/pro/text-to-image
Content updated: 2026-10-01

Plan a product or editorial image from a written brief. Name the subject, framing, light and background separately. This text-to-image endpoint starts from a prompt; choose an editing endpoint when an existing image must be the source.

Suggested starting prompt (not the exact prompt behind the preview):
A ceramic cup on a pale wooden table, viewed at a slight angle. Soft morning light from the left, visible glaze texture, a quiet cream background and space around the cup.

Inputs and settings:
- A text prompt is required. This endpoint does not use a source image as its starting input.
- Choose image_size for the intended layout. Custom dimensions must total between 1,048,576 and 4,194,304 pixels, with an aspect ratio between 1:16 and 16:1; auto_2K is the default preset.
- Start with one image using num_images. Choose JPEG or PNG with output_format, and request a new quote whenever the size or image count changes.

Review the result:
Compare the generated image with your brief. Inspect the rim, handle and shadows at full size. Keep the prompt and settings with your selected output so you can repeat the experiment.

## Segment Anything 2
Source: https://raywake.si/explore/sam2/image
Model ID: sam2/image
Provider documentation: https://fal.ai/models/fal-ai/sam2/image/api
Content updated: 2026-10-01

Select regions in a still image with Segment Anything 2’s image mode. Use spatial selection rather than a scene-generation prompt.

Inputs and settings:
- Send image_url and supply point prompts or box_prompts using the schema’s coordinate structure.
- apply_mask controls applying the selection to the image. For video, choose the separate video operation and its schema.

Review the result:
Check selected boundaries and excluded regions; inspect the output format before relying on transparency.

## SAM 3 Image
Source: https://raywake.si/explore/sam-3/image
Model ID: sam-3/image
Provider documentation: https://fal.ai/models/fal-ai/sam-3/image/api
Content updated: 2026-10-01

Select an object or region in an image to produce segmentation masks. Use a text description, points or boxes to guide the selection; this is not a new-scene generator.

Inputs and settings:
- Provide image_url, then choose prompt, point_prompts or box_prompts for your selection. Check the full schema for coordinate and grouping rules.
- Use include_scores or include_boxes when you need that metadata, and return_multiple_masks when you need separate candidate masks.

Review the result:
Compare each mask with the source at the edges and around overlapping objects. A confidence score is a selection aid, not proof that every pixel is correct.

## SenseNova U1 Infographic
Source: https://raywake.si/explore/sensenova-u1-infographic
Model ID: sensenova-u1-infographic
Provider documentation: https://fal.ai/models/fal-ai/sensenova-u1-infographic/api
Content updated: 2026-10-01

Specify headings, section order, palette and the exact data to display.

Inputs and settings:
- prompt should describe content and layout together. Choose aspect_ratio for the intended slide, poster or vertical graphic; output_format supports JPEG and PNG.
- guidance_scale controls prompt adherence. use_thinking adds a planning pass; review the quote before enabling it. Leave timestep_shift unchanged while refining the content so comparisons remain useful.

Review the result:
Verify every number, label and chart relationship against your source data.

## Stable Audio 2.5
Source: https://raywake.si/explore/stable-audio-25/text-to-audio
Model ID: stable-audio-25/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-25/text-to-audio/api
Content updated: 2026-10-01

Describe instruments, mood and progression, or the physical source of a sound.

Inputs and settings:
- Write prompt and set seconds_total between 1 and 190 seconds. This text-to-audio operation takes no reference recording.
- Raywake accepts 4–8 inference steps and guidance_scale 1–25. Keep seed fixed when comparing one adjustment; judge the complete recording rather than only its opening.

Review the result:
Listen for abrupt endings, unwanted sounds and changes that do not fit the brief.

## Stable Audio 3 Medium
Source: https://raywake.si/explore/stable-audio-3/medium/text-to-audio
Model ID: stable-audio-3/medium/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/medium/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend music and sound with Stable Audio 3 Medium.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 380 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 380 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Audio 3 Small Music Base
Source: https://raywake.si/explore/stable-audio-3/small/music/base/text-to-audio
Model ID: stable-audio-3/small/music/base/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/small/music/base/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend music with Stable Audio 3 Small Music Base.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 120 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 120 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Audio 3 Medium Base
Source: https://raywake.si/explore/stable-audio-3/medium/base/text-to-audio
Model ID: stable-audio-3/medium/base/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/medium/base/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend music and sound with Stable Audio 3 Medium Base.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 380 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 380 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Audio 3 Small Music
Source: https://raywake.si/explore/stable-audio-3/small/music/text-to-audio
Model ID: stable-audio-3/small/music/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/small/music/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend music with Stable Audio 3 Small Music.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 120 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 120 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Audio 3 Small Sfx Base
Source: https://raywake.si/explore/stable-audio-3/small/sfx/base/text-to-audio
Model ID: stable-audio-3/small/sfx/base/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/small/sfx/base/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend sound effects with Stable Audio 3 Small Sfx Base.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 120 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 120 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Audio 3 Small Sfx
Source: https://raywake.si/explore/stable-audio-3/small/sfx/text-to-audio
Model ID: stable-audio-3/small/sfx/text-to-audio
Provider documentation: https://fal.ai/models/fal-ai/stable-audio-3/small/sfx/text-to-audio/api
Content updated: 2026-10-03

Create, transform, repair or extend sound effects with Stable Audio 3 Small Sfx.

Inputs and settings:
- Text mode takes prompt and duration from 1 to 120 seconds. Audio-to-audio also takes an uploaded audio_url and init_noise_level.
- Inpainting replaces only mask_start_seconds to mask_end_seconds; the end must follow the start and stay within the uploaded source.
- Outpainting needs a positive extend_seconds_before or extend_seconds_after. Raywake checks that the source and extensions fit within 120 seconds.
- Choose MP3, WAV, FLAC, OGG, Opus, M4A or AAC. Bitrate applies to compressed formats; base checkpoints expose guidance and more sampling steps.

Review the result:
Listen to the complete output, including edit boundaries, stereo balance, noise and the ending.

## Stable Diffusion 3.5 Medium
Source: https://raywake.si/explore/stable-diffusion-v35-medium
Model ID: stable-diffusion-v35-medium
Provider documentation: https://fal.ai/models/fal-ai/stable-diffusion-v35-medium/api
Content updated: 2026-10-01

Specify subject relationships and lighting for a concept illustration or scene.

Inputs and settings:
- Write prompt; use negative_prompt for exclusions. The provider begins at 40 steps and guidance_scale 4.5; compare adjustments with the same seed.
- Choose 1–4 images with num_images and JPEG or PNG with output_format. auto_fix concerns compatible image sizing; it does not repair visual mistakes in the generated content.

Review the result:
Inspect perspective, facial details and requested text at the actual output size.

## Stable Diffusion 3 Medium
Source: https://raywake.si/explore/stable-diffusion-v3-medium
Model ID: stable-diffusion-v3-medium
Provider documentation: https://fal.ai/models/fal-ai/stable-diffusion-v3-medium/api
Content updated: 2026-10-01

Describe a clear scene and include exact lettering when it is needed.

Inputs and settings:
- Use prompt for the scene and negative_prompt for exclusions. prompt_expansion adds detail; leave it off when comparing your original wording.
- num_images supports 1–4 outputs. The provider starts at 28 steps and guidance_scale 5; change one setting at a time with a recorded seed. No source-image input is exposed here.

Review the result:
Check spelling, object count and relationships between subjects before publishing.

## Stable Diffusion XL
Source: https://raywake.si/explore/fast-sdxl
Model ID: fast-sdxl
Provider documentation: https://fal.ai/models/fal-ai/fast-sdxl/api
Content updated: 2026-10-01

Build illustration or photo-style concepts with a specific subject, setting and light.

Inputs and settings:
- Start with prompt and one image. Compare guidance_scale and num_inference_steps independently; add loras only when you need their effect.
- For source-based changes, choose image-to-image with image_url. Inpainting also requires mask_url; both expose strength. Output format is JPEG or PNG.

Review the result:
Check anatomy, repeated objects and whether a LoRA overwhelms the intended composition.

## SDXL Lightning
Source: https://raywake.si/explore/fast-lightning-sdxl
Model ID: fast-lightning-sdxl
Provider documentation: https://fal.ai/models/fal-ai/fast-lightning-sdxl/api
Content updated: 2026-10-01

Test subject placement and visual direction with a clearly described scene.

Inputs and settings:
- The provider documents 1, 2, 4 or 8 inference steps, with four as the default. Keep a seed when comparing step choices.
- Choose JPEG or PNG in format. Image-to-image adds image_url and strength; inpainting additionally needs mask_url.

Review the result:
Inspect fine edges and recurring detail errors before choosing a version.

## Stable Diffusion with LoRAs
Source: https://raywake.si/explore/lora
Model ID: lora
Provider documentation: https://fal.ai/models/fal-ai/lora/api
Content updated: 2026-10-01

Use a compatible base and adapter for a particular character or style. Image-to-image and inpaint are separate operations.

Inputs and settings:
- The default operation requires model_name and prompt. Add compatible loras and keep the base model fixed when comparing adapter weights.
- Use image_url in image-to-image; supply the source and mask_url for inpaint. noise_strength belongs to those image operations, not the default text mode.

Review the result:
Check subject consistency and unwanted style spillover; inspect repaired edges when using inpaint.

## Stable Video Diffusion
Source: https://raywake.si/explore/fast-svd/text-to-video
Model ID: fast-svd/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/fast-svd/text-to-video/api
Content updated: 2026-10-01

Draft a single-shot motion idea with a stable subject and uncomplicated background.

Inputs and settings:
- The documented output has 25 frames; fps sets playback speed rather than generating more frames. Ten fps produces roughly 2.5 seconds.
- Use motion_bucket_id for motion amount and cond_aug cautiously. Raywake provides steps from 1–100 and negative_prompt for unwanted visual features.

Review the result:
Check the ending frame, subject proportions and visible wobble during motion.

## Stable Video Diffusion LCM
Source: https://raywake.si/explore/fast-svd-lcm/text-to-video
Model ID: fast-svd-lcm/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/fast-svd-lcm/text-to-video/api
Content updated: 2026-10-01

Test subtle subject movement or an atmospheric shot with one coherent scene.

Inputs and settings:
- The provider documents 25 frames. fps changes playback speed: 10 fps yields about 2.5 seconds. Raywake exposes 1–25 fps.
- Adjust motion_bucket_id and cond_aug cautiously. Text mode requires prompt; image mode requires image_url. LCM steps are limited to 1–20.

Review the result:
Watch the whole clip for shape drift, background wobble and unintended movement.

## Storm Trend Effect
Source: https://raywake.si/explore/storm-trend-effect
Model ID: storm-trend-effect
Provider documentation: https://www.eachlabs.ai/eachlabs/eachlabs/storm-trend-effect
Content updated: 2026-10-02

Upload a clear portrait with a visible face for the clip’s central performer.

Inputs and settings:
- Upload the portrait through main_character. This preset takes no prompt or duration control.
- Choose 480p or 720p in resolution, then review the quote. 1080p is unavailable in this release.

Review the result:
Watch the complete clip for changing faces, body proportions and edges during the dance.

## Subway Freeze Video
Source: https://raywake.si/explore/subway-freeze-video
Model ID: subway-freeze-video
Provider documentation: https://www.eachlabs.ai/eachlabs/eachlabs/subway-freeze-video
Content updated: 2026-10-01

Use the preset effect for a character-led clip, optionally featuring a product.

Inputs and settings:
- Supply the person photo in person; product is an optional product image. These fields take images, not a written motion brief.
- Choose resolution from 480p, 720p or 1080p. The current Raywake operation exposes no prompt or duration control; review the resulting timing before placing it in an edit.

Review the result:
Check the face, body outline and product shape throughout the clip.

## T2V Turbo - Video Crafter
Source: https://raywake.si/explore/t2v-turbo
Model ID: t2v-turbo
Provider documentation: https://fal.ai/models/fal-ai/t2v-turbo/api
Content updated: 2026-10-01

Test one simple action and camera direction as a motion sketch.

Inputs and settings:
- Describe the action in prompt. num_frames accepts 16–32 frames and export_fps 1–24; their ratio determines approximate playback length.
- The provider starts with 16 frames at 8 fps, about two seconds. Compare num_inference_steps or guidance_scale separately using seed; changing fps alone changes playback speed rather than adding generated frames.

Review the result:
Watch for flicker, shape changes and an export speed that suits the action.

## Topaz Image Upscale
Source: https://raywake.si/explore/topaz-upscale-image
Model ID: topaz-upscale-image
Content updated: 2026-10-01

Use an existing image when its resolution is too small for your intended layout. Topaz enlarges and enhances that source; it does not create a new scene from a written prompt. Begin with the clearest source you have and keep an untouched copy for comparison.

Inputs and settings:
- Provide the source image using image_url and choose a supported enhancement model. Begin with a modest upscale_factor rather than the largest setting.
- A 2x enlargement doubles both width and height, producing four times the source pixel count. A 4x enlargement produces sixteen times the pixels. Request a quote for the resulting output size.
- Choose output_format and decide whether to enable face enhancement. Check faces against the source before publishing; face recovery can change details as well as sharpen them.

Review the result:
Compare source and output at the same displayed size, then inspect the output at full size. Look at hair, eyes, lettering and straight edges for invented detail, halos or excessive smoothing. Enhanced detail is not proof that missing information was recovered accurately.

## ESRGAN Image Upscaler
Source: https://raywake.si/explore/esrgan
Model ID: esrgan
Provider documentation: https://fal.ai/models/fal-ai/esrgan/api
Content updated: 2026-10-01

Enlarge a source image with a selected RealESRGAN variant. Choose the model for the source type, such as a photograph or anime illustration.

Inputs and settings:
- Provide image_url, select model and choose scale. The Raywake schema accepts scale from 1 to 8 and PNG or JPEG output_format.
- Review face for portraits. tile divides processing into tiles; the provider suggests a nonzero tile size when memory errors occur.

Review the result:
Compare at equal display size, then inspect the enlarged output for halos, altered faces and invented texture.

## Veo 3.1
Source: https://raywake.si/explore/veo3.1
Model ID: veo3.1
Provider documentation: https://fal.ai/models/fal-ai/veo3.1/api
Content updated: 2026-10-01

Use it for a short shot whose subject, action, camera and sound can be described together.

Inputs and settings:
- The default needs prompt. Choose 4s, 6s or 8s, 16:9 or 9:16, and 720p, 1080p or 4k; set generate_audio deliberately.
- Image mode needs image_url; first/last mode needs first_frame_url and last_frame_url. Reference mode uses image_urls and has no duration control in Raywake.

Review the result:
Review dialogue, lip movement and whether the shot follows the intended action.

## Veo 3.1
Source: https://raywake.si/explore/veo3.1/extend-video
Model ID: veo3.1/extend-video
Provider documentation: https://fal.ai/models/fal-ai/veo3.1/extend-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Extend a 720p or 1080p source in 16:9 or 9:16. The new segment is seven seconds at 720p. generate_audio changes the price; check the quote after choosing audio.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Veo 3.1 Fast
Source: https://raywake.si/explore/veo-3-1-fast
Model ID: veo3.1/fast
Content updated: 2026-10-01

Turn a short scene description into a moving sequence. Describe the subject, action and camera movement separately; the jungle example keeps attention on one animal and its surroundings.

Suggested starting prompt (not the exact prompt behind the preview):
A jaguar slowly turns toward the camera in a lush jungle. Gentle movement in the leaves, soft daylight through the canopy and a steady close camera.

Inputs and settings:
- The featured endpoint starts with a text prompt. If you need to animate a particular photo, select the image-to-video mode and its corresponding model ID.
- Choose a landscape 16:9 or vertical 9:16 frame with aspect_ratio. Select duration and resolution from the current schema; describe an action that fits the chosen length.
- Decide whether to include audio with generate_audio. Describe intended dialogue or ambience, then review the actual soundtrack rather than assuming every requested sound will be accurate.

Review the result:
Play with sound enabled when reviewing your result. Check the action from start to finish and confirm that any audio suits the scene before publishing.

### Is the Veo 3.1 Fast API free on Raywake?
Generating with Veo 3.1 Fast uses Raywake credits. You can request a quote before submitting; obtaining the quote does not start a generation or charge credits. Any free tier offered by a different service has its own terms.
Quotes and credits: https://raywake.si/docs/credits

## Veo 3.1 Fast
Source: https://raywake.si/explore/veo3.1/fast/extend-video
Model ID: veo3.1/fast/extend-video
Provider documentation: https://fal.ai/models/fal-ai/veo3.1/fast/extend-video/api
Content updated: 2026-10-02

Select the operation for your source, upload any required media, and describe the result you want.

Inputs and settings:
- Extend a 720p or 1080p source in 16:9 or 9:16. The new segment is seven seconds at 720p. generate_audio changes the price; check the quote after choosing audio.

Review the result:
Watch the entire clip and compare motion, identity, transitions and audio with your inputs.

## Veo 3.1 Lite
Source: https://raywake.si/explore/veo3.1/lite
Model ID: veo3.1/lite
Provider documentation: https://fal.ai/models/fal-ai/veo3.1/lite/api
Content updated: 2026-10-01

Use it for a scene draft with one readable action before refining the shot.

Inputs and settings:
- For the default text operation, provide prompt and choose 4s, 6s or 8s, 16:9 or 9:16, and 720p or 1080p.
- The image operation additionally requires image_url and offers auto aspect ratio. Use negative_prompt for exclusions; auto_fix may rewrite a rejected prompt.

Review the result:
Check the framing, action timing and unwanted sounds before using the clip.

## Bria Video Background Removal
Source: https://raywake.si/explore/bria/video/background-removal
Model ID: bria/video/background-removal
Provider documentation: https://fal.ai/models/bria/video/background-removal/api
Content updated: 2026-10-01

Replace a clip’s background with the original Bria video-removal operation. Keep the source footage available when judging changes near moving edges.

Inputs and settings:
- Provide video_url. The provider specifies a source smaller than 4000x4000 and shorter than 30 seconds for this operation.
- Choose background_color and output_container_and_codec. Set preserve_audio; for transparency, use a compatible codec and verify the file in your editor.

Review the result:
Check hands, hair and object crossings from start to finish. Verify that the selected background and audio settings match the export.

## Bria Green Screen Despill
Source: https://raywake.si/explore/bria/video/background-removal/green-screen-despill
Model ID: bria/video/background-removal/green-screen-despill
Provider documentation: https://fal.ai/models/bria/video/background-removal/green-screen-despill/api
Content updated: 2026-10-01

Clean up a green-screen clip with Bria’s despill operation. Use footage actually filmed against a green screen rather than an arbitrary background.

Inputs and settings:
- Provide video_url. spill_suppression ranges from 0 to 1 in Raywake; 0 disables suppression.
- Choose output_container_and_codec and preserve_audio. Confirm alpha support in the selected codec and test the result on the intended background.

Review the result:
Inspect green fringes, clothing colors and hair across frames. Compare suppressed edges against the original before compositing.

## Bria Video Erase
Source: https://raywake.si/explore/bria/video/erase/prompt
Model ID: bria/video/erase/prompt
Provider documentation: https://fal.ai/models/bria/video/erase/prompt/api
Content updated: 2026-10-01

Clean a shot by naming the unwanted subject; use mask mode when a visual selection is more precise.

Inputs and settings:
- Prompt mode requires video_url and prompt. Mask mode instead requires video_url and mask_video_url.
- Use footage under 5 seconds, as specified by the provider. Review auto_trim, preserve_audio and output_container_and_codec so trimming and export match your edit.

Review the result:
Inspect the cleared area across every frame for flicker, ghosting and altered neighboring objects.

## Vidu Q2 Pro
Source: https://raywake.si/explore/vidu/q2/image-to-video/pro
Model ID: vidu/q2/image-to-video/pro
Provider documentation: https://fal.ai/models/fal-ai/vidu/q2/image-to-video/pro/api
Content updated: 2026-10-01

Use it when an existing composition should lead into a planned motion or final frame.

Inputs and settings:
- Provide image_url and a prompt up to 3,000 characters. Select 2–8 seconds and 720p or 1080p; adjust movement_amplitude and optional end_image_url.
- Video extension instead requires video_url, offers 2–7 seconds and an optional prompt. Compare changes with the same seed and inspect the ending before extending.

Review the result:
Check subject shape through the transition and continuity at any video-extension join.

## Vidu Q2 Turbo
Source: https://raywake.si/explore/vidu/q2/image-to-video/turbo
Model ID: vidu/q2/image-to-video/turbo
Provider documentation: https://fal.ai/models/fal-ai/vidu/q2/image-to-video/turbo/api
Content updated: 2026-10-01

Use it to animate a prepared scene while keeping the desired opening composition explicit.

Inputs and settings:
- image_url and prompt are required; prompt accepts up to 3,000 characters. Choose 2–8 seconds and 720p or 1080p.
- Add end_image_url for a planned ending. Describe the path between frames, adjust movement_amplitude, and keep seed unchanged when comparing motion directions.

Review the result:
Watch for drifting facial features and abrupt movement as the clip approaches its ending image.

## Vidu Q2 Image
Source: https://raywake.si/explore/vidu/q2/text-to-image
Model ID: vidu/q2/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/vidu/q2/text-to-image/api
Content updated: 2026-10-01

Use it for a scene concept, or switch to reference mode when an existing subject should inform the image.

Inputs and settings:
- The default takes prompt, up to 1,500 characters. Select 16:9, 9:16 or 1:1 and describe placement, lighting and material details.
- Reference mode additionally requires reference_image_urls. Explain which features to retain and what should change; use seed to compare variations under matching settings.

Review the result:
Check subject count, small objects and any lettering before treating the image as a finished asset.

## Vidu Q2
Source: https://raywake.si/explore/vidu/q2/text-to-video
Model ID: vidu/q2/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/vidu/q2/text-to-video/api
Content updated: 2026-10-01

Use it to draft a shot without a starting image, describing subject movement and camera movement separately.

Inputs and settings:
- Provide prompt, up to 3,000 characters. Choose 2–8 seconds and 16:9, 9:16 or 1:1 according to the planned placement.
- Resolution options are 360p, 520p, 720p and 1080p. Adjust movement_amplitude for the subject; retain seed when comparing a revised scene description.

Review the result:
Check that camera motion does not obscure the action and that objects remain recognizable across frames.

## Vidu Reference to Image
Source: https://raywake.si/explore/vidu/reference-to-image
Model ID: vidu/reference-to-image
Provider documentation: https://fal.ai/models/fal-ai/vidu/reference-to-image/api
Content updated: 2026-10-01

Use it to place a referenced character or object in a newly described scene.

Inputs and settings:
- Provide prompt and reference_image_urls. Keep the description within 1,500 characters and identify which subject each reference should contribute.
- Choose 16:9, 9:16 or 1:1. State the new setting and features that should remain; hold seed fixed while testing changes to composition.

Review the result:
Compare identifying features, clothing and proportions against the references; inspect overlaps between subjects.

## Vidu Q1
Source: https://raywake.si/explore/vidu/q1/text-to-video
Model ID: vidu/q1/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/vidu/q1/text-to-video/api
Content updated: 2026-10-01

Use it to explore a general or anime scene with a specific amount of subject movement.

Inputs and settings:
- The default requires prompt, up to 1,500 characters. Choose general or anime style and 16:9, 9:16 or 1:1; adjust movement_amplitude.
- Image mode needs image_url; reference mode needs reference_image_urls. Start/end mode needs start_image_url and end_image_url. Keep seed fixed when comparing prompt edits.

Review the result:
Inspect limb motion and transitions; compare reference subjects against their source images.

## WAN 2.2 5B Image
Source: https://raywake.si/explore/wan/v2.2-5b/text-to-image
Model ID: wan/v2.2-5b/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.2-5b/text-to-image/api
Content updated: 2026-10-01

Explore portrait lighting, materials or a single scene composition.

Inputs and settings:
- prompt is the required input. Describe one composition and use negative_prompt for unwanted elements; this mode takes no source image.
- Compare guidance_scale and num_inference_steps one at a time with a fixed seed. Decide whether enable_prompt_expansion should elaborate your brief.

Review the result:
Inspect skin, textures and object edges at full size; verify any lettering separately.

## Wan 2.2 5B Fast
Source: https://raywake.si/explore/wan/v2.2-5b/text-to-video/fast-wan
Model ID: wan/v2.2-5b/text-to-video/fast-wan
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.2-5b/text-to-video/fast-wan/api
Content updated: 2026-10-01

Try one subject action and one camera move before building a longer sequence.

Inputs and settings:
- Write prompt as a shot description. Choose aspect_ratio from 16:9, 9:16 or 1:1 and resolution from 480p, 580p or 720p.
- Keep seed fixed while comparing guidance_scale. Review motion before changing interpolator_model; Raywake exposes no duration control for this mode.

Review the result:
Check motion direction, temporal flicker and whether the ending completes the action.

## WAN 2.2 A14B Image
Source: https://raywake.si/explore/wan/v2.2-a14b/text-to-image
Model ID: wan/v2.2-a14b/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.2-a14b/text-to-image/api
Content updated: 2026-10-01

Develop a scene from a brief or revise an existing composition in image-to-image mode.

Inputs and settings:
- Default mode requires prompt. Compare guidance_scale and guidance_scale_2 separately with the same seed and brief.
- For image-to-image, supply image_url and prompt. Use strength between 0 and 1 to balance preservation against remaking; select that mode before adding image controls.

Review the result:
Inspect fine textures and geometry; for edits, compare the preserved subject with the source.

## Wan 2.2 A14B Turbo
Source: https://raywake.si/explore/wan/v2.2-a14b/text-to-video/turbo
Model ID: wan/v2.2-a14b/text-to-video/turbo
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.2-a14b/text-to-video/turbo/api
Content updated: 2026-10-01

Use a concise shot brief; switch modes when an existing composition should anchor the clip.

Inputs and settings:
- Default mode requires prompt. Choose aspect_ratio and resolution; Raywake does not expose duration in this mode.
- Image-to-video requires image_url and prompt, with optional end_image_url. Choose a compatible ending image and describe the movement connecting the two frames.

Review the result:
Watch subject identity, camera movement and unwanted cropping; review the last frame.

## WAN 2.7 Pro Image
Source: https://raywake.si/explore/wan/v2.7/pro/text-to-image
Model ID: wan/v2.7/pro/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.7/pro/text-to-image/api
Content updated: 2026-10-01

Develop a product or editorial scene with explicit material, lighting and object placement.

Inputs and settings:
- Text mode requires prompt, capped at 2000 characters. num_images supports 1–5; choose output_format as jpeg, png or webp.
- Editing requires image_urls and prompt and permits 1–4 outputs. Keep seed fixed and put unwanted content in negative_prompt, capped at 500 characters.

Review the result:
Inspect reflections, small lettering and object proportions; compare edits with every reference.

## WAN 2.7 Image
Source: https://raywake.si/explore/wan/v2.7/text-to-image
Model ID: wan/v2.7/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.7/text-to-image/api
Content updated: 2026-10-01

Describe the main subject and spatial relationships before requesting variations.

Inputs and settings:
- Default mode needs prompt up to 2000 characters. Use num_images for 1–5 outputs and select jpeg, png or webp with output_format.
- For editing, provide image_urls and prompt; num_images is limited to 1–4. Save seed for comparisons and keep negative_prompt within 500 characters.

Review the result:
Check object counts, layout and lettering at delivery size; compare preserved details after editing.

## WAN 2.5 Image
Source: https://raywake.si/explore/wan-25-preview/text-to-image
Model ID: wan-25-preview/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/wan-25-preview/text-to-image/api
Content updated: 2026-10-01

Use it to explore visual alternatives or edit a scene while describing the features that should remain.

Inputs and settings:
- For text generation, enter prompt, use negative_prompt for exclusions, and choose num_images from 1 to 4. Compare alternatives with a fixed seed.
- The editing operation requires image_urls and prompt. Keep changes specific. Raywake does not expose image_size here; default prompt expansion is separate from editing controls.

Review the result:
Check text, object counts and whether edits unintentionally changed the main subject.

## Wan 2.5
Source: https://raywake.si/explore/wan-25-preview/text-to-video
Model ID: wan-25-preview/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/wan-25-preview/text-to-video/api
Content updated: 2026-10-01

Use it for a scene paced around one action or a supplied soundtrack.

Inputs and settings:
- The default requires prompt. Select 5 or 10 seconds, 16:9, 9:16 or 1:1, and 480p, 720p or 1080p.
- audio_url can guide the soundtrack; enable_prompt_expansion may rewrite the description. Image mode additionally requires image_url and does not expose aspect_ratio.

Review the result:
Listen for audio cutoffs and check that the action completes within the selected duration.

## Wan 2.6 Uncensored
Source: https://raywake.si/explore/wan-2-6-uncensored
Model ID: wan-2-6-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-2-6-uncensored
Content updated: 2026-10-01

Animate a still or pace a text-led scene with an audio track.

Inputs and settings:
- Enter prompt, up to 2,000 characters; choose 5, 10 or 15 seconds and 720P or 1080P. negativePrompt allows up to 500 characters.
- inputImage and inputAudio each accept one file. An image determines framing; ratio applies to text-only work. Set audioEnabled, promptExtend and shotType for the intended clip.

Review the result:
Check motion continuity and listen for a silent tail if the supplied audio is shorter than the video.

## Wan 2.7 Image
Source: https://raywake.si/explore/wan-2-7-image
Model ID: wan-2-7-image
Provider documentation: https://wiro.ai/models/alibaba/wan-2-7-image
Content updated: 2026-10-01

Use it for product composites or edits with clearly assigned reference roles.

Inputs and settings:
- Provide prompt, up to 5,000 characters. inputImage accepts up to nine images; refer to them by order when describing what to retain or combine.
- Raywake currently limits samples to 1. Choose the available size and keep seed fixed for comparisons; do not request the larger image sets advertised upstream.

Review the result:
Check that the product shape and requested edits are correct, especially where images are blended.

## Wan 2.7 Uncensored
Source: https://raywake.si/explore/wan-2-7-uncensored
Model ID: wan-2-7-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-2-7-uncensored
Content updated: 2026-10-01

Use it when a prepared portrait, illustration or product image should define the opening frame.

Inputs and settings:
- Provide exactly one image in inputImage and a nonempty prompt. Choose 5, 10 or 15 seconds and 720p or 1080p.
- The source image sets the aspect ratio. inputAudio accepts one optional track; use negativePrompt for exclusions and set audioEnabled or promptExtend according to the shot.

Review the result:
Review facial changes, fine edges and how the subject moves away from the starting pose.

## Wan 3.0
Source: https://raywake.si/explore/alibaba/wan-3.0/image-to-video
Model ID: alibaba/wan-3.0/image-to-video
Provider documentation: https://fal.ai/models/alibaba/wan-3.0/image-to-video/api
Content updated: 2026-10-01

Animate a starting image with Wan 3.0. Describe movement within the existing frame; optionally supply a compatible ending image for a planned transition.

Inputs and settings:
- start_image_url is required; end_image_url is optional. Choose duration, resolution and aspect_ratio before requesting a quote.
- prompt describes motion. Record enable_prompt_expansion when comparing instructions, and use audio to choose whether to include generated sound.

Review the result:
Compare the first and last frames with your sources. Watch for identity changes, drifting backgrounds and an abrupt ending; listen to generated sound when enabled.

## Wan 3.0
Source: https://raywake.si/explore/wan-3-0
Model ID: wan-3-0
Provider documentation: https://wiro.ai/models/alibaba/wan-3-0
Content updated: 2026-10-01

Use it for a product reveal or narrative shot whose opening and ending are already planned.

Inputs and settings:
- Provide prompt and select 5–30 seconds in 5-second steps. Raywake offers 480P, 720P and 1080P; choose ratio for the intended framing.
- inputImage and inputImageLast each accept one image. Describe the movement between them, set audioEnabled for sound, and review any promptExtend rewrite.

Review the result:
Check that the action reaches its destination without sudden shape changes or skipped motion.

## Wan 3.0 Prime
Source: https://raywake.si/explore/alibaba/wan-3.0-prime/image-to-video
Model ID: alibaba/wan-3.0-prime/image-to-video
Provider documentation: https://fal.ai/models/alibaba/wan-3.0-prime/image-to-video/api
Content updated: 2026-10-01

Turn an established composition into a moving shot; describe motion instead of rebuilding the whole scene.

Inputs and settings:
- Provide start_image_url. Use optional prompt for movement and end_image_url for a target ending frame.
- Choose 2–30 seconds, 480p/720p/1080p and aspect_ratio, including adaptive. Set audio for sound; compare prompt-expansion changes with your intended action.

Review the result:
Check subject identity, camera movement and continuity into any supplied ending frame.

## Wan 3.0 Prime
Source: https://raywake.si/explore/wan-3-0-prime
Model ID: wan-3-0-prime
Provider documentation: https://wiro.ai/models/alibaba/wan-3-0-prime
Content updated: 2026-10-01

Use it to test a camera move between two planned compositions before assembling a longer edit.

Inputs and settings:
- Enter prompt, choose 5, 10, 15, 20, 25 or 30 seconds, and select 480P, 720P or 1080P in Raywake.
- Add one image each to inputImage and inputImageLast when needed. Set ratio and audioEnabled explicitly; retain seed while adjusting only the camera direction.

Review the result:
Compare the final pose with the ending image and check that background details stay coherent.

## Wan 3.0 Prime Pro Uncensored
Source: https://raywake.si/explore/wan-3-0-prime-pro-uncensored
Model ID: wan-3-0-prime-pro-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-3-0-prime-pro-uncensored
Content updated: 2026-10-01

Use it for a framed product shot or character scene with a defined final composition.

Inputs and settings:
- Provide prompt. Choose 5–30 seconds in 5-second steps and 1080p, 2k or 4k; set ratio to match the intended placement.
- inputImage and inputImageLast each accept one optional frame. Specify the connecting action and sound, choose audioEnabled, and keep seed fixed while refining promptExtend or the prompt.

Review the result:
Inspect small textures at full size and ensure the ending frame is reached without an abrupt jump.

## Wan 3.0 Prime Uncensored
Source: https://raywake.si/explore/wan-3-0-prime-uncensored
Model ID: wan-3-0-prime-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-3-0-prime-uncensored
Content updated: 2026-10-01

Use it for a before-and-after transition with a clear visual start and destination.

Inputs and settings:
- Enter prompt and choose 5, 10, 15, 20, 25 or 30 seconds. Resolution choices are 480p, 720p and 1080p.
- Use inputImage and inputImageLast for one starting and ending image each. Describe how the change unfolds; set ratio, audioEnabled and promptExtend to suit that transition.

Review the result:
Check the middle of the transition for changing proportions and listen for unexpected dialogue or effects.

## Wan 3.0 Pro Uncensored
Source: https://raywake.si/explore/wan-3-0-pro-uncensored
Model ID: wan-3-0-pro-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-3-0-pro-uncensored
Content updated: 2026-10-01

Use it for a continuous scene where camera movement, lighting and an ending pose are important.

Inputs and settings:
- Write prompt for one continuous shot, then choose 5–30 seconds in 5-second steps and 1080p, 2k or 4k.
- inputImage and inputImageLast each accept one frame. Specify camera direction and sound; set ratio and audioEnabled. Keep seed unchanged when revising a single detail.

Review the result:
Inspect fine textures across moving frames, including at higher output tiers.

## Wan 3.0 Uncensored
Source: https://raywake.si/explore/wan-3-0-uncensored
Model ID: wan-3-0-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/wan-3-0-uncensored
Content updated: 2026-10-01

Use it to turn character art into a motion test with a specific starting pose.

Inputs and settings:
- Provide prompt and select 5, 10, 15, 20, 25 or 30 seconds. Available resolutions are 480p, 720p and 1080p.
- Each of inputImage and inputImageLast accepts one image. Describe the action between poses, select ratio and audioEnabled, and review the result if promptExtend rewrites the brief.

Review the result:
Check identity, clothing and edge shimmer as the character moves; compare the final composition with any ending image.

## Wan 2.7
Source: https://raywake.si/explore/wan/v2.7/text-to-video
Model ID: wan/v2.7/text-to-video
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.7/text-to-video/api
Content updated: 2026-10-01

Use a scene brief for a new shot or a source frame when the starting composition matters.

Inputs and settings:
- Text mode requires prompt. Choose duration from 2–15 seconds, resolution as 720p or 1080p and aspect_ratio for the layout.
- In image mode, use image_url with optional end_image_url, or video_url for continuation. Do not combine image_url and video_url; audio_url is optional.

Review the result:
Check motion continuity, the ending transition and sound alignment across the clip.

## Wan VACE 14B
Source: https://raywake.si/explore/wan-vace-14b
Model ID: wan-vace-14b
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/api
Content updated: 2026-10-01

Use it for a localized video edit or a reframing task where reference material should guide the scene.

Inputs and settings:
- Provide prompt and select task for the edit. Add video_url, ref_image_urls or endpoint frames as needed; align masks with the source.
- Choose 480p, 580p or 720p. Set num_frames and frames_per_second together for pacing; mask_image_url takes precedence over mask_video_url when both are supplied.

Review the result:
Inspect mask boundaries through motion and ensure the edit does not alter areas you intended to preserve.

## Wan VACE 14B Depth
Source: https://raywake.si/explore/wan-vace-14b/depth
Model ID: wan-vace-14b/depth
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/depth/api
Content updated: 2026-10-01

Use depth guidance to explore a new appearance around an existing scene layout.

Inputs and settings:
- Supply video_url and prompt. Check whether your guidance clip needs preprocess before adding reference images through ref_image_urls.
- Choose resolution from 480p, 580p or 720p. Keep seed fixed while comparing guidance_scale; inspect motion before adding num_interpolated_frames.

Review the result:
Compare foreground silhouettes, depth boundaries and camera movement throughout the clip.

## Wan VACE 14B Inpainting
Source: https://raywake.si/explore/wan-vace-14b/inpainting
Model ID: wan-vace-14b/inpainting
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/inpainting/api
Content updated: 2026-10-01

Use a mask to replace a selected object or repair a bounded part of a shot.

Inputs and settings:
- Provide video_url and prompt with mask_video_url or mask_image_url. Use one mask source to avoid ambiguous precedence.
- Align mask and source framing. Select resolution and keep seed fixed when changing guidance_scale; inspect the entire edited sequence.

Review the result:
Watch mask boundaries and moving objects; check that nearby regions remain suitable.

## Wan VACE 14B Outpainting
Source: https://raywake.si/explore/wan-vace-14b/outpainting
Model ID: wan-vace-14b/outpainting
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/outpainting/api
Content updated: 2026-10-01

Extend a tight shot when the surrounding scene needs more room.

Inputs and settings:
- Provide video_url and prompt describing the surrounding scene. Select the needed expand_left, expand_right, expand_top or expand_bottom sides.
- Use expand_ratio conservatively within 0–1. Match aspect_ratio to delivery and compare added edges across the whole clip.

Review the result:
Check seams, perspective and moving objects crossing the original frame boundary.

## Wan VACE 14B Pose
Source: https://raywake.si/explore/wan-vace-14b/pose
Model ID: wan-vace-14b/pose
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/pose/api
Content updated: 2026-10-01

Explore a character or costume while using an existing movement sequence as guidance.

Inputs and settings:
- Supply video_url and prompt. Check preprocess for your guidance source; ref_image_urls can provide appearance references.
- Choose resolution and frames_per_second for the shot. Keep seed fixed while refining guidance_scale; avoid adding interpolation until the motion is usable.

Review the result:
Compare limb positions, foot contact and timing; check identity consistency between frames.

## Wan VACE 14B Reframe
Source: https://raywake.si/explore/wan-vace-14b/reframe
Model ID: wan-vace-14b/reframe
Provider documentation: https://fal.ai/models/fal-ai/wan-vace-14b/reframe/api
Content updated: 2026-10-01

Prepare a shot for a different layout while monitoring the subject’s position.

Inputs and settings:
- video_url is required; prompt is optional. Set aspect_ratio for the intended layout and inspect the source before enabling trim_borders.
- zoom_factor crops more of the edges as it increases. Start near zero and compare the full shot before adding first_frame_url or last_frame_url references.

Review the result:
Check lost edge detail, subject visibility and regenerated borders throughout the shot.

## Wan v2.2 A14B Text-to-Image A14B with LoRAs
Source: https://raywake.si/explore/wan/v2.2-a14b/text-to-image/lora
Model ID: wan/v2.2-a14b/text-to-image/lora
Provider documentation: https://fal.ai/models/fal-ai/wan/v2.2-a14b/text-to-image/lora/api
Content updated: 2026-10-03

Generate still images with WAN 2.2 A14B and optional LoRA weights.

Inputs and settings:
- Set prompt, image_size, negative_prompt and seed. Each LoRA has a path, scale and high/low transformer selection.
- Choose regular or no acceleration, guidance settings and 2–40 sampling steps. PNG and JPEG outputs are supported.

Review the result:
Compare the result with your source, especially object boundaries, fine text, faces and shadows.

## Wan v2.6 Image to Image
Source: https://raywake.si/explore/wan/v2.6/image-to-image
Model ID: wan/v2.6/image-to-image
Provider documentation: https://fal.ai/models/wan/v2.6/image-to-image/api
Content updated: 2026-10-01

Assign a clear role to each reference, such as subject, background or visual style.

Inputs and settings:
- Provide image_urls and prompt. Official guidance supports 1–3 references; refer to their order as image 1, image 2 and image 3.
- Start num_images at 1; Raywake allows up to 4. Decide whether enable_prompt_expansion should rewrite the instruction and keep seed fixed for comparisons.

Review the result:
Compare subject identity, lighting compatibility and compositing edges with the references.

## WAN 2.6 Image
Source: https://raywake.si/explore/wan/v2.6/text-to-image
Model ID: wan/v2.6/text-to-image
Provider documentation: https://fal.ai/models/wan/v2.6/text-to-image/api
Content updated: 2026-10-02

Describe the image or edit you want and upload any required source images.

Inputs and settings:
- Use the current input form for this endpoint, then review the credit quote before generating.

Review the result:
Inspect the full-size output and compare details and subject appearance with your source.

## Wan 2.6
Source: https://raywake.si/explore/wan/v2.6/text-to-video
Model ID: wan/v2.6/text-to-video
Provider documentation: https://fal.ai/models/wan/v2.6/text-to-video/api
Content updated: 2026-10-01

Plan a continuous shot or a short scene sequence with explicit timing.

Inputs and settings:
- Text mode requires prompt; image mode also requires image_url. Choose duration as 5, 10 or 15 seconds and resolution as 720p or 1080p.
- multi_shots needs enable_prompt_expansion enabled. If using audio_url, match its length to the clip: longer audio is cut and shorter audio leaves silence.

Review the result:
Check scene transitions, subject continuity and sound coverage through the ending.

## Wan 2.1
Source: https://raywake.si/explore/wan-t2v
Model ID: wan-t2v
Provider documentation: https://fal.ai/models/fal-ai/wan-t2v/api
Content updated: 2026-10-01

Use it to draft a motion scene, then switch operations if the opening composition needs a source image.

Inputs and settings:
- The default requires prompt and offers 16:9 or 9:16. Use negative_prompt for unwanted elements and hold seed fixed when comparing descriptions.
- Image mode adds image_url; first/last mode needs start_image_url and end_image_url. These modes expose 480p/720p, guide_scale and shift; default text mode does not.

Review the result:
Check that the subject moves as described and that details remain consistent between endpoint frames.

## Workflow Utilities Blend Video
Source: https://raywake.si/explore/workflow-utilities/blend-video
Model ID: workflow-utilities/blend-video
Provider documentation: https://fal.ai/models/fal-ai/workflow-utilities/blend-video/api
Content updated: 2026-10-01

Combine two clips as overlapping layers with Blend Video. Plan which clip is the top layer and which supplies the background.

Inputs and settings:
- Provide top_video_url and bottom_video_url. Choose blend_mode and set opacity from 0 to 1 for the top layer.
- With shortest enabled, output ends when the shorter source ends. Check source durations before requesting the composite.

Review the result:
Review the entire composite for visibility, timing and an abrupt ending. This blends frames rather than placing clips one after another.

## Workflow Utilities Extract Nth Frame
Source: https://raywake.si/explore/workflow-utilities/extract-nth-frame
Model ID: workflow-utilities/extract-nth-frame
Provider documentation: https://fal.ai/models/fal-ai/workflow-utilities/extract-nth-frame/api
Content updated: 2026-10-01

Extract regularly spaced still frames from an existing video for review or another image workflow. This operation samples frames; it does not animate a still image.

Inputs and settings:
- Provide video_url. frame_interval counts frames, not seconds: 12 selects every twelfth frame.
- Limit output with max_frames. Choose output_format; quality controls JPEG/WebP output. Request a quote for the complete selection.

Review the result:
Check the returned frame count and inspect the selected frames. A regular interval can miss a brief action; choose the interval for the timing of your source video.

## Workflow Utilities Reverse Video
Source: https://raywake.si/explore/workflow-utilities/reverse-video
Model ID: workflow-utilities/reverse-video
Provider documentation: https://fal.ai/models/fal-ai/workflow-utilities/reverse-video/api
Content updated: 2026-10-01

Reverse an existing clip’s playback with the video utility. Choose source footage whose backwards motion suits the intended sequence.

Inputs and settings:
- Provide video_url; no prompt is required.
- The provider reverses both video and audio streams. This operation changes playback order rather than generating new scene content.

Review the result:
Watch the opening, ending and reversed soundtrack. Spoken words or a musical beat may no longer fit the scene.

## Workflow Utilities Scale Video
Source: https://raywake.si/explore/workflow-utilities/scale-video
Model ID: workflow-utilities/scale-video
Provider documentation: https://fal.ai/models/fal-ai/workflow-utilities/scale-video/api
Content updated: 2026-10-01

Resize an existing video for a different layout. Choose whether to retain the whole frame, crop the edges or stretch the image.

Inputs and settings:
- Provide video_url and at least width or height, using even pixel dimensions. One dimension preserves the aspect ratio; two dimensions use mode.
- pad retains the frame with borders; crop fills the frame by center-cropping; stretch may distort proportions. Select codec, preset and crf for encoding.

Review the result:
Play the resized export. Check subject framing, distorted proportions and visible compression; resizing alone does not restore missing detail.

## Z Image Base
Source: https://raywake.si/explore/z-image/base
Model ID: z-image/base
Provider documentation: https://fal.ai/models/fal-ai/z-image/base/api
Content updated: 2026-10-01

Explore a scene’s materials and lighting while adjusting prompt adherence.

Inputs and settings:
- Supply prompt and use negative_prompt for exclusions. Keep seed fixed while testing guidance_scale and num_inference_steps independently.
- Start num_images at 1; Raywake permits up to 4. Choose output_format as jpeg, png or webp, and leave acceleration unchanged during comparisons.

Review the result:
Compare texture, geometry and unwanted elements at full size; higher settings need visual review.

## Z Image Pro Uncensored
Source: https://raywake.si/explore/z-image-pro-uncensored
Model ID: z-image-pro-uncensored
Provider documentation: https://wiro.ai/models/wiro-partners/z-image-pro-uncensored
Content updated: 2026-10-01

Develop a character or editorial scene with explicit framing and lighting.

Inputs and settings:
- Provide prompt; choose resolution as 1k or 2k and ratio for framing. Raywake uses ratio, not aspect_ratio, for this operation.
- promptExtend is a string: use "true" or "false" rather than a boolean. Keep seed within 0–2147483647 and save it when comparing prompt changes.

Review the result:
Inspect anatomy, fabric edges and any lettering at full size.

## Z-Image Turbo
Source: https://raywake.si/explore/z-image/turbo
Model ID: z-image/turbo
Provider documentation: https://fal.ai/models/fal-ai/z-image/turbo/api
Content updated: 2026-10-01

Explore a new scene or switch modes to revise an existing photo.

Inputs and settings:
- Text mode requires prompt. num_inference_steps supports 1–8, num_images supports 1–4 and output_format offers jpeg, png or webp.
- Image-to-image needs image_url and prompt; inpaint also needs mask_image_url. Both expose strength above 0 through 1; adjust it separately from the text brief.

Review the result:
Inspect fine detail; for masked edits, check the boundary and nearby untouched areas.

## GPT Image 1
Source: https://raywake.si/explore/gpt-image-1/text-to-image
Model ID: gpt-image-1/text-to-image
Provider documentation: https://fal.ai/models/fal-ai/gpt-image-1/text-to-image/api
Content updated: 2026-10-01

Build a written composition, then use editing when an existing image must guide the result.

Inputs and settings:
- Text mode requires prompt; image_size includes auto and three fixed sizes. quality and num_images control output choices.
- Edit mode requires image_urls and supports input_fidelity. Use PNG or WebP when requesting a transparent background.

Review the result:
Compare spelling, lighting and object proportions with the brief or reference.