Introduction
A practical guide to using the Remove Background model (Bria) in our Media Generator: options, costs, and best practices.
Remove Background (bria/remove-background) cuts the main subject from its background, producing a transparent or clean-background image. In our platform, it supports:
- Background removal — Isolate subjects for composites, catalogs, or ads
- PNG output — Supports transparency
- Reference image required — One input image
- Low cost — Flat per-image pricing
Use it for product cutouts, headshots, and any image where you need the subject isolated.
Where to Use Remove Background
| Location | How | |----------|-----| | Media Generator | Tools → Media Generator → Create Job → choose "Remove Background" → upload a reference image | | Image projects | Create a project with reference images and select Remove Background |
Requirements
- Reference image required — One input image
- Output aspect ratio follows the input (matchinputimage)
- Output format: PNG only (for transparency)
Platform Options
Aspect ratio
- matchinputimage only — Output keeps the input image’s aspect ratio
Output format
- PNG only — Required for transparent background
Reference images
- Required — Exactly one image
- The model segments the subject and removes the rest
Costs (credits)
Remove Background uses flat per-image pricing:
| Item | USD | Credits | |------|-----|----------| | Per image | $0.02 | 0.05 |
\Credits computed with default billing config. Check the model selector in the UI for current rates.
Prompt guidelines
- The prompt has limited effect; the model focuses on background removal based on the image.
- You can optionally add context (e.g., “product photo,” “person portrait”) if helpful.
Good practices
- Use clear subject contrast — Subjects that stand out from the background work best.
- Prefer simple backgrounds — Cleaner backgrounds usually give cleaner cuts.
- High-resolution input — Better input quality improves edge quality.
- Check edges — Inspect hair, fine details, and transparent areas.
Limits and constraints
| Constraint | Value | |------------|-------| | Reference images | Required, 1 per generation | | Max images per request | 5 | | Output format | PNG only | | Aspect ratio | matchinputimage only |
Troubleshooting
| Issue | What to try | |-------|-------------| | Missing reference warning | Upload an image to process | | Rough or inaccurate edges | Use higher resolution; improve subject/background contrast | | Partial removal | Ensure the subject is clearly defined in the frame | | Model not available | Confirm Replicate API token is configured |
