Models and service identity
During early access, we release performance improvements and updates based on your feedback as needed during our daily 05:00–06:00 UTC update window. Releases will become less frequent as we approach the end of early access.
GET /v1/models
Call https://api.blatent.ai/v1/models with your API key. The current response is:
{"models":[{"name":"blatent-struct-v0.0.1a","description":"JSON Schema constrained generation.","release_date":"2026-09-28"}]}
The optional model in a resolve request must be blatent-struct-v0.0.1a when non-null. Omit it to use the deployed model. Other names return 422.
Every response has an X-Blatent-Request-ID header for request correlation. Retain it with errors and timings, and include it when you contact support.
Health checks
| Endpoint | Authentication | Successful response |
|---|---|---|
GET /v1/health | None | 200 with an empty body: the API is reachable. |
GET /v1/health/auth | Bearer API key | 200 with an empty body: the key is valid. |
An absent or invalid key on /v1/health/auth returns 401. Neither endpoint runs inference, checks available credit or proves that a resolve request will succeed. Do not try to parse their successful responses as JSON.
This version supports JSON Schema generation with optional state and OPA policies. Use resolve to generate an answer and the schema registry to reuse definitions.