API Reference

The Robotoff API provides programmatic access to Open Food Facts' AI-powered product data extraction and processing capabilities. This comprehensive API enables developers to integrate advanced computer vision, machine learning, and data processing features into their applications.

Authentication

Most endpoints are publicly accessible. For operations that require authentication, you'll need to provide appropriate credentials through the Open Food Facts authentication system.

API Categories

The Robotoff API is organized into several functional categories:

Core Prediction Services

• Prediction Management: Manage AI predictions, retrieve prediction results, and handle prediction workflows
• Image Predictions: Process product images and extract structured data using computer vision
• Predict Services: Direct prediction endpoints for text analysis, language detection, and category prediction

Content Management

• Logo Management: Detect, annotate, and manage product logos and brand marks
• Annotation Management: Handle human annotations and crowdsourced data validation
• Question Management: Manage validation questions and community feedback workflows

Image Processing

• Image Management: Upload, process, and manage product images for analysis
• Image Processing: Advanced image processing operations and OCR capabilities

Data & Insights

• Insight Management: Manage AI-generated insights and data extraction results
• Dataset Management: Manage training datasets and model data

Operations & Monitoring

• Batch Processing: Execute and monitor large-scale batch processing operations
• User Management: User statistics and contribution tracking
• ANN Search: Approximate nearest neighbor search for similar products
• System Status: Health checks, API status, and system monitoring

Getting Started

Common Use Cases

Product Data Extraction

• Extract ingredient lists from product packaging
• Identify nutritional information tables
• Detect product categories and classifications
• Extract brand information and logos

Quality Assurance

• Validate existing product data
• Generate improvement suggestions
• Crowdsource data validation through questions
• Track annotation quality and user contributions

Batch Processing

• Process large volumes of product images
• Update existing product information
• Generate insights for data analysis
• Export processed data for further analysis

Research and Development

• Access training datasets
• Experiment with different prediction models
• Analyze prediction accuracy and performance
• Contribute to model improvements

Response Format

All API responses follow a consistent JSON structure:

{
  "status": "success",
  "data": {
    // Response data
  },
  "meta": {
    "count": 100,
    "page": 1,
    "pages": 10
  }
}

Error Handling

The API uses standard HTTP status codes and provides detailed error messages:

• 200 - Success
• 400 - Bad Request (invalid parameters)
• 401 - Unauthorized (authentication required)
• 404 - Not Found (endpoint or resource not found)
• 429 - Rate Limited (too many requests)
• 500 - Internal Server Error

Rate Limits

API requests are subject to rate limiting to ensure fair usage:

• General endpoints: 100 requests per minute
• Image processing: 20 requests per minute
• Batch operations: 5 requests per minute

SDKs and Libraries

While there are no official SDKs, the API follows REST conventions and can be easily integrated using standard HTTP libraries in any programming language:

• Python: requests, httpx
• JavaScript: fetch, axios
• Java: OkHttp, HttpClient
• PHP: Guzzle, cURL

Support and Community

• GitHub Issues: Report bugs and request features
• Slack Community: Join the Open Food Facts Slack workspace
• API Updates: Follow the changelog for API changes