Explore the documentation
Start
Activate your API key, send your first chat completion with an official OpenAI SDK or curl, and learn the conventions every request shares.
Getting started →Chat
OpenAI-compatible chat completions with streaming, citations, function calling, structured output, server-side conversations, models, and embeddings.
Chat completions →Knowledge base
Upload files or crawl your public website, manage the documents the assistant answers from, and track ingestion jobs with signed webhooks.
Ingesting files →Reliability
One machine-readable error envelope, per-account rate limits with Retry-After, key handling and data isolation, and a stable /v1 versioning commitment.
Errors and reliability →Built for production
OpenAI-compatible
Point the official OpenAI SDKs at the base URL and call chat completions, models, and embeddings as you already do. SDKs
Grounded answers with citations
Answers are drawn from your knowledge base and carry a citations array naming the passages they used. Citations
Your knowledge base
Upload files or crawl your public site; ingestion runs as jobs that you poll or receive signed webhooks for. Jobs and webhooks
Errors you can program against
Every error is one envelope, {"error": {...}}, with a stable code, a type, and the offending param. Errors
Resources
- Interactive API reference: every endpoint, request, and response, with a try-it console.
- OpenAPI specification: the machine-readable contract, also published as YAML; generate clients in your language.
- AI agents: a copy-paste onboarding prompt, plus llms.txt for tools.
- eCourtDate.com: access requires a paid subscription; API keys are issued and activated for your account by eCourtDate.