Odock Documentation

Build governed AI products faster.

Odock sits between your applications and model or MCP providers, giving teams one place to manage access, safety, budgets, routing, usage, and observability.

Open docs Docker quickstart

Choose your path

The most useful starting points for evaluating, integrating, and operating Odock.

Start with the platform

Understand how Odock governs LLM and MCP traffic across the gateway, UI, and observability stack.

Run it locally

Bring up Odock with Docker Compose and get the core services ready for development.

Use the LLM gateway

Send OpenAI-compatible, Anthropic, Gemini, vLLM, and unified chat requests through one governed endpoint.

Control access

Configure API keys, providers, models, MCP servers, teams, and policy boundaries from the control plane.

Set budgets and quotas

Reserve, settle, and enforce spend or usage limits across organisations, teams, users, and API keys.

Operate with visibility

Track logs, metrics, traces, alerts, dashboards, request identity, usage, latency, and cost.