Programmatic access: GET /api/changelog returns this data as JSON. Every response includes an X-API-Version header with the current version date.
v2026-03-02March 2, 2026Latest
- GET /api/usage now accepts Authorization: Bearer <api_key> header auth
- Query api_key auth on /api/usage is deprecated fallback for migration only
- Docs updated with preferred header-based examples and migration guidance
v2026-02-28February 28, 2026
- Added X-API-Version header to all responses
- Added /v1/changelog endpoint for programmatic version tracking
- Added X-RateLimit-* headers (Limit, Remaining, Reset) to all authenticated endpoints
- Added webhook management API: CRUD, test ping, delivery history
- Added onboarding checklist API for first-time user guidance
- Phase 2 parameter normalization: dpi, page_range, detect_tables support
- File upload validation: 20MB maximum file size enforced client-side
v2026-02-27February 27, 2026
- Added country of origin metadata to engine configs
- Engine detail endpoints: /v1/engines/:id/stats, /v1/engines/:id/latency, /v1/engines/:id/uptime
- Phase 1 parameter normalization: unified settings.language with automatic code mapping
- Native engine passthrough via native_settings parameter
- Response includes applied_settings showing resolved configuration
v2026-02-25February 25, 2026
- Enabled cloud OCR engines: openai/gpt-4o, openai/gpt-4.1-mini-ocr, anthropic/claude-sonnet, google/gemini-2.5-flash
- Fixed connection pooling: switched to session-mode for prepared statement support
- Added /v1/engines/health endpoint for real-time engine availability
- Added docTR, RapidOCR, OCRmyPDF to local engine fleet
v2026-02-22February 22, 2026
- Initial public API release (v1)
- Core endpoint: POST /v1/ocr with base64 and URL input types
- Batch processing: POST /v1/ocr/batch (up to 20 items)
- 6 local OCR engines: Tesseract, EasyOCR, docTR, Surya, RapidOCR, OCRmyPDF
- API key authentication with SHA-256 hashed storage
- Usage tracking and Stripe-based billing
- Idempotency-Key support for safe retries