The Xenition API is live
· Xenition
Today we shipped the public Xenition API. You can now build applications using chat, embeddings, image, audio, and video generation endpoints, with usage tracking and a redesigned Developers page.
API improvements
- API keys now have a per-minute request limit
- Failed API calls are never charged
- POST requests support Idempotency-Key to prevent duplicate charges
- Xenition public model names are available for API calls; GET /v1/models lists them
New endpoints
- POST /chat/completions with sync and streaming, charged only on success
- POST /embeddings supporting float and base64 formats
- POST /images/generations for image generation
- POST /audio/speech and /audio/transcriptions endpoints
- Async video generation endpoint with signed webhook notifications
Developer tools
- Usage tracking, per-call cost lookup, and developer request log
- /app/developers page for managing API keys, usage, request log, and webhook secret
- Access built-in tools via the API with the new /tools and /tools/{id}/run endpoints
Another 19 smaller fixes and improvements shipped today — the full list is on the changelog.