Available now

Bria Extract Object: object extraction

Bria Extract Object cuts out the object you name in a photo and returns it on its own, with everything else removed.

Model ID

Input

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

Source image as a public, a base64 (data:image/png;base64,...) or raw base64-encoded image bytes.

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

Natural-language description of the object to act on, e.g. 'the red car'.

Example prompt. Edit it or write your own.

Additional settings

Tighten the output canvas to the extracted object.

Estimated total$0.04

You sign up first. Your settings are kept.

About this model

Overview

Upload an image and describe the object in a few words, such as "the red car". The model finds that object and returns a cutout. Autocrop trims the canvas to the object, and a background removal pass can refine the edges.

What you can build

  • One product cut out from a busy group photo
  • Assets lifted from a scene for a new layout
  • Logos or packaging isolated for a mockup

Prompting tips

  • Describe the object by color and position when the image holds several similar ones, for example "the blue mug on the right".
  • Leave background removal off for text, logos and graphics; it can trim thin parts of them.
  • Turn on autocrop when you want a tight file with no empty space around the object.

What to check

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

Inputs

  • Image (required): Source image as a public URL, a base64 data URI (data:image/png;base64,...) or raw base64-encoded image bytes.
  • Prompt (required): Natural-language description of the object to act on, e.g. 'the red car'.
  • Autocrop (optional): Tighten the output canvas to the extracted object.
  • Remove Background (optional): When True, refine the cutout alpha with background removal (RMBG). When False (default), use the SAM segmentation mask as the cutout's alpha -- faster, with no salient-object matting (helpful when RMBG over-trims text, logos, or graphics).

Output and limits

  • 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

  • 1 image: $0.04 / image