Available now

Ideogram Character: character inpainting

By Ideogram

Ideogram Character keeps one person's look from a single reference photo and places them in new scenes, edits or remixes.

Model ID
Maker
Ideogram

Input

Fields marked * are required. Review the price before generating.

A set of images to use as style references (maximum total size 10MB across all style references). The images should be in JPEG, PNG or WebP format

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

The prompt to fill the masked part of the image.

Example prompt. Edit it or write your own.

MUST have the exact same dimensions (width and height) as the mask image.

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

The mask to inpaint the image. MUST have the exact same dimensions (width and height) as the input image.

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

A set of images to use as character references. Currently only 1 image is supported, rest will be ignored. (maximum total size 10MB across all character references). The images should be in JPEG, PNG or WebP format

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

A set of masks to apply to the character references. Currently only 1 mask is supported, rest will be ignored. (maximum total size 10MB across all character references). The masks should be in JPEG, PNG or WebP format

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

Additional settings

A list of 8-character hexadecimal codes for the style of the image. Cannot be used with style reference images.

Determine if MagicPrompt should be used in generating the request or not.

Seed for the random number generator

Estimated total · Balanced$0.24

You sign up first. Your settings are kept.

About this model

Overview

Give one reference image of a character and describe a scene; the face and distinctive features carry over into each new image. Up to 8 images per run, in square, portrait or landscape sizes, with Turbo, Balanced or Quality rendering.

The edit mode paints the character into a masked area of an image you already have. The remix mode reworks an existing image around the character, with a strength setting for how much of the original stays.

What you can build

  • A recurring mascot or host across a series of posts
  • Storyboards and comic panels with the same lead
  • One person shown in many settings for a campaign
  • A character added into an existing photo or layout

Prompting tips

  • Use one clear, well lit photo with the face fully visible; only the first reference is used.
  • Describe the scene, outfit and action in the prompt; the reference handles who the person is.
  • In the edit mode the image and mask must have exactly the same width and height.
  • In the remix mode lower the strength to move further from the original image.

What to check

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

Inputs

  • Images (optional): A set of images to use as style references (maximum total size 10MB across all style references). The images should be in JPEG, PNG or WebP format.
  • Rendering Speed (optional): The rendering speed to use. Options: Turbo, Balanced, Quality.
  • Style Codes (optional): A list of 8-character hexadecimal codes for the style of the image. Cannot be used with style reference images.
  • Enable Prompt Expansion (optional): Determine if MagicPrompt should be used in generating the request or not.
  • Number of Images (optional): Number of images to generate. From 1 to 8.
  • Seed (optional): Seed for the random number generator.
  • Prompt (required): The prompt to fill the masked part of the image.
  • Image (required): The image URL to generate an image from. MUST have the exact same dimensions (width and height) as the mask image.
  • Mask (required): The mask URL to inpaint the image. MUST have the exact same dimensions (width and height) as the input image.
  • Reference Images (required): A set of images to use as character references. Currently only 1 image is supported, rest will be ignored. (maximum total size 10MB across all character references). The images should be in JPEG, PNG or WebP format.
  • Reference Masks (optional): A set of masks to apply to the character references. Currently only 1 mask is supported, rest will be ignored. (maximum total size 10MB across all character references). The masks should be in JPEG, PNG or WebP format.

Output and limits

  • Up to 8 images per run.
  • Small or dense text in the image can come out misspelled.
  • Hands, crowds and complex poses can show artifacts; generate a few variations and pick the best.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • Turbo: $0.16 / image
  • Balanced: $0.24 / image
  • Quality: $0.32 / image