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.

# Bria Image Editing API v2 – Quickstart Guide

Unlock Bria’s generative editing power by fine-tuning core parameters to match your use case. This guide walks you through best practices, recommended settings, and tips for reproducible, high-quality outputs.

## 1. Core Parameter Overview
- **image** (required): URL or Base64 string of your source.  
- **aspect_ratio**: Choose from `1:1`, `16:9`, `9:16`, etc.  
- **canvas_size** [width, height]: Defines overall output size in pixels.  
- **original_image_size** [w, h]: Controls the input image’s footprint on the canvas.  
- **original_image_location** [x, y]: Offsets the image within the canvas.  
- **prompt / negative_prompt**: Guides generative fill or background replacement.  
- **seed**: Any integer to ensure consistency across runs.  
- **preserve_alpha**: Keep transparency when working with PNGs.  
- **prompt_content_moderation**, **visual_input_content_moderation**, **visual_output_content_moderation**: Auto-filter NSFW content.

## 2. Best Practices
1. **Lock Your Seed**  
   Use `seed=12345` (or any fixed integer) for reproducible edits.  
2. **Moderate Content**  
   Enable `prompt_content_moderation=true` and visual moderation flags for safe deployments.  
3. **Mask Precision**  
   For Eraser and Fill operations, supply clean binary masks to isolate the edit region.  
4. **Leverage Negative Prompts**  
   Exclude unwanted elements (e.g., `"Do not include buildings"`) to refine generative outputs.

## 3. Parameter Tuning by Use Case

### E-commerce & Product Photography
- aspect_ratio: `1:1`  
- Remove Background: `preserve_alpha=false`  
- Increase Resolution: `canvas_size=[2000,2000]`, `seed=42`  
- prompt: _unused_ / negative_prompt to avoid shadows.

### Marketing & Advertising Banners
- aspect_ratio: `16:9`  
- Replace Background: `prompt="vibrant city skyline at dusk"`  
- canvas_size: `[1920,1080]`, original_image_location:`[100,50]`  
- seed: `2023`

### Digital Art & Concept Sketches
- Expand Image: `canvas_size=[3000,2000]`  
- original_image_size:[1000,1000], original_image_location:[500,500]  
- prompt: `"mystical forest with glowing mushrooms"`  
- negative_prompt: `"modern buildings"`

### Social Media Portrait Touch-Ups
- aspect_ratio: `9:16` or `2:3`  
- Generative Fill: `prompt="soft skin smoothing, warm lighting"`  
- negative_prompt: `"blemishes, harsh shadows"`  
- preserve_alpha: `false` for JPEG-style export

### UI/UX Prototypes & Transparent Assets
- preserve_alpha: `true`  
- canvas_size: match screen dimensions (e.g., `[1440,1024]`)  
- prompt: `"transparent background"`  
- seed: fixed for consistent mockups

## 4. Pro Tips
- Always preview low-res drafts before final upscaling.  
- Combine `visual_output_content_moderation=true` with brand safety workflows.  
- Use `original_image_size` & `location` to composite multiple elements precisely.  

By tailoring these parameters, you’ll harness Bria Image Editing API v2 to produce professional, pixel-perfect results across any visual workflow.

FAQs