LangChain just rolled out OpenWiki, an open source CLI built to make documentation more accessible for AI coding agents. Instead of the usual documents crafted for developers to skim, this tool builds a structured wiki that is tailor made for large language models (LLMs). This gives your coding agents a clearer view of how your repository is set up, how files connect, and the coding patterns before they make any edits.
The tool follows Andrej Karpathy’s “LLM Wiki” idea, which suggests that instead of relying on retrieval augmented generation (RAG) every time you ask a question, you keep a living wiki that an LLM updates and maintains. According to LangChain, the goal is to reduce the effort required to keep documentation current while giving coding agents durable context that leads to more informed code changes and fewer avoidable mistakes. The tool creates an openwiki/ directory into your repository, links it to coding agents with AGENTS.md or CLAUDE.md, and keeps everything updated as your project evolves.
What is OpenWiki and Why does it Matter?
Most repositories already contain documentation such as READMEs, docstrings, and architecture diagrams. The catch is that those are built for humans, not AI agents working with limited context windows. As those repositories get bigger and more people push changes, documentation can also become outdated. That makes it tough for coding agents to really grasp what is new or changed.
This tool addresses this challenge by generating a structured Markdown wiki designed specifically for AI consumption. It relies on summaries, clear headings, and cross references so agents can be focused on what they need fast.
The tool manages AGENTS.md or CLAUDE.md by adding in instructions that tell coding agents to check out the generated wiki when they need context. If those files are not there yet, OpenWiki creates them. LangChain explains that this approach avoids placing extensive documentation directly inside instruction files, allowing agents to retrieve only the documentation they need rather than loading hundreds of pages into every session.
To keep documents from falling out of date, the tool offers a GitHub Action you can schedule once daily. This workflow scans commits since the last run, checks Git diffs for repository changes, updates the wiki where needed, and opens a pull request with the new documentation. That means less manual maintenance and more reliability.
This is actually useful.
— Min Choi (@minchoi) July 3, 2026
LangChain just released OpenWiki.
It's an open-source agent that creates a wiki for your codebase, connects it to your coding agent, and keeps it updated as your repo changes.
Your AI coding agent gets long-term repo context without stuffing everything… pic.twitter.com/fPkg6D3edM
How to Use OpenWiki
GitHub explains that getting started begins by installing the CLI using:
| npm install -g openwiki |
After installation, users initialize the tool by running:
| openwiki –init |
During your first interactive session, OpenWiki walks you through configuring an inference provider, your API key, and the LLM you want to use. It saves those details locally in:
| ~/.openwiki/.env |
If you want, you can add a LangSmith API key to track the tool’s activity in a LangSmith project called “openwiki”.
After setup, you are set to generate documentation with the interactive CLI (just type openwiki), start it with a prompt like openwiki “Please generate documentation for this repository” or start off a single command using the -p option. To refresh documents, go with openwiki –update. Need a rundown of commands? openwiki –help covers them.
To automate maintenance, users can copy the provided:
| openwiki-update.yml file into .github/workflows/openwiki-update.yml in their repository. |
The schedule starts off the tool, auto updates documentation, and files a pull request with any fresh documentation. The tool works out of the box with OpenRouter, Fireworks, Baseten, OpenAI, and Anthropic. You can also tweak custom model IDs for the supported inference providers.
Also read: Meta Quietly Introduces Pocket – Here’s What It Is, & How It Works
OpenWiki takes the effort out of keeping codebase docs current by generating an AI ready wiki that coding agents can rely on for context. Building on the LLM Wiki model, it offers structured documentation and keeps everything in sync through scheduled GitHub updates. Developers spend less time maintaining documents, while coding agents get up to date context that makes their work smarter and safer. The tool is all about keeping documentation locked in-step with your code without forcing developers to go through endless manual updates.
