Documentation
CiviX API documentation
Everything you need to call the CiviX Inference API: authentication, streaming, models, and the error contract.
Get started
API reference
Chat Completions
POST /v1/chat/completions — parameters, responses, and multimodal input.
Streaming
Server-sent events, keepalive comments, and why long prompts should stream.
Errors
A single error shape, stable machine-readable codes, and what to retry.
Rate limits & capacity
Why you might get a 429, and how to behave when you do.