What will live here
An interactive reference (powered by Stoplight Elements) covering the public endpoints:
| Area | Endpoint | Purpose |
|---|---|---|
| Chat | POST /v1/chat/completions | Route a request to a model |
| Prompts | GET /v1/prompts/{name} | Resolve the active prompt version |
| Usage | GET /v1/usage | Query structured usage records |
| Budgets | GET /v1/budgets | Inspect budget consumption |
| Audit | GET /v1/audit | Read audit events |
Each entry will include request/response schemas, authentication details, examples in curl / Go / TypeScript, and a try-it-out console for sandbox calls.
Why we are not publishing the spec yet
The OpenAPI file in our backend repository currently describes the internal surface (/internal/v1/*, governance, admin). Publishing it on the marketing site would be misleading — what you can actually call as a design partner is the OpenAI-compatible chat endpoint, and the schema we publish here needs to reflect that contract, not the internal one.
Once the public surface is finalized in the upcoming Phase, we will publish a schneeai-public.yaml here and the reference will render automatically.
Get notified
Email [email protected] with “API Reference” in the subject line and we will let you know the moment it goes live. If you are already a design partner and need specific endpoint details before then, reach out via your usual channel.
For today, the developer docs and token cost calculator cover everything you need to evaluate and integrate.