AI Documentation Generator

AI Documentation Generator analyzes repositories and writes documentation for developers who need to understand an unfamiliar codebase or prepare it for new contributors. It's open source under the MIT license and runs as a Python application, in Docker, or through a Claude Code plugin.

Its analysis splits the work among specialized AI agents that examine code structure, dependencies, data movement, request handling, and APIs in parallel. It saves their findings as reusable documents, then uses them to produce a README and project instructions for coding assistants. Those outputs include CLAUDE.md, AGENTS.md, and Cursor rules, so the same code analysis can inform both developer documentation and AI assistant context.

You choose the model backend. The standalone application supports local models through an OpenAI-compatible API, as well as OpenAI, OpenRouter, and Anthropic-compatible gateways. Different agents can use different models or endpoints. The application runs on your own machine or server; choosing a cloud endpoint sends model requests to that provider. Generated documents stay in the repository.

README generation can use an existing README as context, and you can choose which sections to include. Assistant-rule generation can preserve existing files and vary the level of detail. For teams using GitLab, scheduled analysis finds recently active projects and opens merge requests with updated findings. A Helm chart supports scheduled deployment on Kubernetes.