Bria Guided Remove Background | [image-to-image]
Bria Guided Remove Background delivers prompt-controlled background removal at $0.03 per image, allowing users to precisely define which objects to keep or remove using natural language. Starting from a standard background removal cutout, the model lets users include, exclude, or isolate specific elements through simple text instructions.
Built on commercially licensed training data, the model produces transparent cutouts without regenerating or repainting image content. Only the alpha channel is modified, preserving the original colors, textures, and details of retained objects. Designed for teams that need precise, commercially cleared image segmentation with flexible subject selection.
Use Cases: E-commerce Product Photography | Marketing Asset Production | Selective Object Extraction | Professional Photo Editing
At $0.03 per image, Bria Guided Remove Background provides prompt-controlled background removal, allowing users to include, exclude, or isolate specific objects while preserving the original image pixels.
| Metric | Result | Context |
|---|
| Training Data | 100% licensed commercial imagery | Cleared for commercial use |
| Cost per Image | $0.03 | Approximately 33 generations per $1.00 on fal |
| Input | Image + natural-language prompt | Text instructions for subject selection |
| Output Format | PNG with transparency | Original image pixels preserved |
| Prompt Control | Include, exclude, or isolate objects | Supports multiple objects in a single prompt |
| Image Processing | Alpha-channel modification | No regeneration or repainting of image content |
Built for Precise, Prompt-Guided Background Removal
Bria Guided Remove Background uses natural-language instructions to control which elements remain visible in an image. Users can specify multiple objects, exclude unwanted elements, or isolate a single subject without creating masks, clicking on objects, or using specialized prompt syntax.
What this means for you:
- Natural-language object selection: Define what to keep or remove using simple prompts such as "only the chair" or "the person without the lamp."
- Multi-object segmentation: Retain multiple subjects and surrounding elements in a single cutout, including people, products, and objects within a scene.
- Original pixel preservation: Only the alpha channel changes, keeping the original colors, textures, and details of retained objects untouched.
- Transparent PNG output: Generate cutouts ready for compositing, product listings, and professional image editing workflows.
- Commercially licensed training data: Built on Bria's models, trained exclusively on licensed data and cleared for commercial use.
- Straightforward integration: Use natural-language prompts to control background removal through an API, without manual masking or object selection.
The same source image can produce different cutouts depending on the intended application. For example, a home-decor brand can retain a person, yoga mat, and plant; a wellness app can keep the person and mat; and a yoga-wear brand can isolate just the person. Each result preserves the original image pixels while adjusting which elements remain visible.
Technical Specifications
| Spec | Details |
|---|
| Model | Bria Guided Remove Background |
| Task | Prompt-guided background removal |
| Input | Image and natural-language prompt |
| Output Format | PNG with alpha channel transparency |
| Object Selection | Prompt-guided inclusion, exclusion, and isolation |
| Image Preservation | Original image pixels retained; only alpha channel modified |
| Image Regeneration | None |
| License | Commercial use with licensed training data |
| Pricing | $0.03 per image |
Quickstart Guide | Enterprise Pricing
Designed for Flexible Image Editing
Bria Guided Remove Background is designed for photographs with identifiable subjects, including people, products, interiors, and multi-object scenes. Its prompt-guided approach enables users to adapt the same source image for different creative and commercial requirements without modifying the underlying image content.
When a prompt specifies the obvious primary subject, the model produces the same clean cutout as standard background removal. If a prompt references an object that is not present in the image, the model falls back to the default background removal result, retaining the primary subject.
Because the model modifies transparency rather than reconstructing image content, regions previously hidden behind removed objects remain transparent instead of being artificially filled. For example, removing a mug from someone's hand leaves the previously occluded area transparent rather than generating new hand details.
No prompt needed? Bria Remove Background automatically isolates the main subject at $0.018 per image.
bria/guided-remove-background
