rmbg-api

REST API server for removing backgrounds from images using AI models.

Features

• 🚀 Fast background removal using ONNX Runtime
• 🔒 Privacy-focused - all processing done server-side (no third-party uploads)
• 📦 Multiple model support (u2netp, modnet, briaai, and more)
• 🐳 Docker support for easy deployment
• 📊 Health check and model listing endpoints

Quick Start

Using Docker

# Build the Docker image
docker build -t rmbg-api -f packages/api/Dockerfile .

# Run the container
docker run -p 3000:3000 rmbg-api

Using Node.js

# Install dependencies
pnpm install

# Build the API
pnpm --filter rmbg-api build

# Start the server
pnpm --filter rmbg-api start

# Or run in development mode
pnpm --filter rmbg-api dev

API Endpoints

Health Check

GET /health

Response:

{
  "status": "ok",
  "timestamp": "2024-01-01T00:00:00.000Z"
}

List Models

GET /models

Response:

{
  "models": [
    {
      "name": "u2netp",
      "resolution": 320,
      "size": 4574861
    },
    ...
  ]
}

Remove Background

POST /remove-background
Content-Type: multipart/form-data

image: [file]
model: "u2netp" (optional, default: "u2netp")
maxResolution: 2048 (optional, default: 2048)

Example using cURL:

curl -X POST http://localhost:3000/remove-background \
  -F "[email protected]" \
  -F "model=u2netp" \
  -F "maxResolution=2048" \
  --output output.png

Example using fetch:

const formData = new FormData()
formData.append('image', file)
formData.append('model', 'u2netp')
formData.append('maxResolution', '2048')

const response = await fetch('http://localhost:3000/remove-background', {
  method: 'POST',
  body: formData
})

const blob = await response.blob()

Available Models

• u2netp - 320px, 4.5MB (fastest, default)
• modnet - 512px, 25MB (medium quality)
• briaai - 1024px, 44MB (highest quality)
• isnet-anime - 1024px, 168MB (anime-optimized)
• silueta - 320px, 43MB (specialized)
• u2net-cloth - 768px, 170MB (clothing-focused)

Environment Variables

• PORT - Server port (default: 3000)
• NODE_ENV - Environment mode (production/development)

Error Handling

The API returns appropriate HTTP status codes and JSON error messages:

• 400 - Bad request (invalid model, missing file, etc.)
• 413 - File too large (max 10MB)
• 500 - Internal server error

Example error response:

{
  "error": "Invalid model",
  "message": "Model \"invalid\" not found. Available models: u2netp, modnet, briaai"
}

Docker Deployment

Using docker-compose

Create docker-compose.yml:

version: '3.8'
services:
  api:
    build:
      context: .
      dockerfile: packages/api/Dockerfile
    ports:
      - "3000:3000"
    environment:
      - PORT=3000
      - NODE_ENV=production
    restart: unless-stopped

Run:

docker-compose up -d

Production Considerations

1. Resource Limits: Set appropriate CPU/memory limits
2. Load Balancing: Use multiple instances behind a load balancer
3. Caching: Consider caching model downloads
4. Monitoring: Implement logging and monitoring solutions
5. Rate Limiting: Add rate limiting middleware for production use

License

See root LICENSE file.