OpenWiki is a LangChain-maintained CLI that turns codebase documentation into a living context layer for AI coding agents. Its core promise is straightforward: generate an openwiki/ directory for a repository, keep that documentation updated as the code changes, and make it easy for agents to reference the wiki rather than overloading AGENTS.md, CLAUDE.md or a single giant prompt file. The tool fits a common developer workflow problem: coding agents can write code quickly, but they often lose architectural context, repeat stale assumptions, or waste tokens rediscovering the same repo structure. OpenWiki frames documentation as agent infrastructure rather than a static human-only artifact.
The live GitHub source is strong enough for a create-ready tool page. As of the July 5 write check, langchain-ai/openwiki was an active MIT-licensed repository with 4,392+ GitHub stars, 326+ forks, recent pushes, and a README that documents npm install -g openwiki, openwiki --init, interactive usage, one-shot -p mode, and openwiki --update. The README also explains the daily GitHub Action path for opening documentation update PRs and notes that OpenWiki can append prompting to AGENTS.md or CLAUDE.md so a coding agent knows where to search for context. It supports OpenRouter, Fireworks, Baseten, OpenAI and Anthropic out of the box, with custom model IDs available through configuration.
The aicoolies page should position OpenWiki as a documentation-and-context tool for teams already adopting Claude Code-style agents, Cursor, Codex, or other repository agents. Strong use cases include onboarding a large repo, keeping architecture notes fresh, giving agents a durable search target, and reducing the temptation to paste huge context files into every session. Caveats should mention that generated wiki quality depends on model choice, repo structure and review discipline; secrets are saved locally in ~/.openwiki/.env; and generated documentation still needs human review before it becomes trusted engineering knowledge.
