Bria Image Editing API v2 – Generative Image Editing Model

What is Bria Image Editing API v2?

Bria Image Editing API v2 is a suite of generative AI–powered image manipulation models designed for developers, creators, and product managers. Whether you need selective object removal, background transformation, canvas expansion, or high-fidelity upscaling, Bria delivers pixel-accurate edits while preserving original fidelity and alpha transparency. It integrates seamlessly via REST endpoints, accepts common image URIs or Base64 inputs, and supports advanced parameters—from aspect_ratio to negative_prompt—to fine-tune results. The model is exclusively trained on licensed data, ensuring safe and risk-free commercial use.

Key Features

• Eraser: Remove unwanted objects or blemishes using precise mask inputs.
• Generative Fill: Add new elements guided by text prompts within masked regions.
• Remove Background (RMBG 2.0): Automatically isolate and eliminate backgrounds with refined edge detection.
• Replace Background: Swap backgrounds for solid colors or AI-generated scenes by supplying a prompt.
• Expand Image: Extend canvas size in any direction, using contextual prompts to fill new areas.
• Enhance Image: Sharpen details, reduce artifacts, and boost clarity without altering composition.
• Increase Resolution: Upscale images while preserving texture and avoiding regeneration artifacts.
• Transparent Alpha Support: Maintain PNG transparency throughout editing pipelines.

Best Use Cases

• E-commerce & Product Photos: Cleanly remove backgrounds, enhance detail, and upsample for high-resolution catalogs.
• Marketing & Advertising: Quickly create fresh backgrounds or composite elements for banners and social posts.
• Digital Art & Design: Sketch expansions and generative fills accelerate concept iterations.
• Social Media Content: Auto-remove distracting elements and retouch portraits on the fly.
• UI/UX Prototypes: Generate seamless assets with transparent alpha for mockups.

Prompt Tips and Output Quality

• Use prompt and negative_prompt fields to steer generative fill and background replacement.
• Specify aspect_ratio, canvas_size, original_image_size, and original_image_location for precise layout control.
• Set a seed for reproducible results across runs.
• Enable preserve_alpha to keep transparent regions intact (ideal for PNGs).
• Tweak prompt_content_moderation to automatically filter NSFW terms.
• For artifact-free upscaling, rely on the Increase Resolution model rather than re-rendering.

FAQs

Q: How do I remove an object from my photo?
A: Use the Eraser model with a binary mask. Supply your image URI and mask via the image and mask parameters. The API returns a clean composite without the object.

Q: Can I change my photo’s background to a custom scene?
A: Yes. Use Replace Background and include a text prompt describing your desired scene. For solid colors, set the prompt to the color name (e.g., “solid light gray”).

Q: What input formats are supported?
A: Bria accepts public image URLs or Base64-encoded strings. Transparent PNGs keep their alpha channel when preserve_alpha is true.

Q: How do I expand my image canvas and fill it?
A: Call Expand Image with canvas_size larger than the original, and include a descriptive prompt (e.g., “sunset over rolling hills”) for context-aware generation.

Q: How do I ensure consistent outputs?
A: Provide a fixed integer seed value. Identical seeds with the same parameters yield reproducible edits.