> ## Documentation Index
> Fetch the complete documentation index at: https://docs.langchain.com/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenWiki

> CLI that writes and maintains agent wikis so coding agents work faster

OpenWiki is an open source CLI that writes and maintains a Markdown wiki about your codebase or personal knowledge. The wiki captures details such as architecture, integrations, evals, and workflows so [coding agents](/oss/python/deepagents/overview) can use it as durable context instead of rediscovering the repository on every task.

That makes agent work faster and cheaper in tokens: agents read a curated wiki first, then inspect source only where they need more detail. Humans can browse the same Markdown (and the local [visualizer](/oss/openwiki/visualize)), but the primary audience is agents.

OpenWiki is built on [Deep Agents](/oss/python/deepagents/overview) and supports tracing with [LangSmith](/langsmith/observability-quickstart).

## Get started

Install the CLI, then initialize documentation for the current repository:

```bash theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
npm install -g openwiki
openwiki --init
```

See the [Quickstart](/oss/openwiki/quickstart) to choose a model provider, generate docs, and keep them up to date.

## Modes

OpenWiki has two modes:

| Mode               | Command                      | Output                                | Use when                                                        |
| ------------------ | ---------------------------- | ------------------------------------- | --------------------------------------------------------------- |
| **Code** (default) | `openwiki` / `openwiki code` | `openwiki/` in the current repository | You want repository context and documentation for coding agents |
| **Personal**       | `openwiki personal`          | `~/.openwiki/wiki`                    | You want a local personal brain from configured sources         |

Bare `openwiki --init` and `openwiki --update` run in code mode. Use `openwiki personal --init` or `openwiki personal --update` for the personal wiki.

## Capabilities

<CardGroup cols={2}>
  <Card title="Repository wikis" icon="folder-code" href="/oss/openwiki/code-mode">
    Generate Markdown docs under `openwiki/`, then wire them into `AGENTS.md` and `CLAUDE.md` so coding agents can find them.
  </Card>

  <Card title="Personal brain" icon="brain" href="/oss/openwiki/personal-mode">
    Build a local wiki from git repos, Gmail, Notion, web search, Hacker News, and X/Twitter.
  </Card>

  <Card title="Automatic updates" icon="clock" href="/oss/openwiki/automate-updates">
    Refresh docs from GitHub Actions, GitLab CI, or Bitbucket Pipelines and open a PR when content changes.
  </Card>

  <Card title="Model providers" icon="cpu" href="/oss/openwiki/providers">
    Use OpenAI, Anthropic, Gemini, Bedrock, OpenRouter, GitHub Copilot, and other providers out of the box.
  </Card>

  <Card title="Open Knowledge Format" icon="file-text" href="/oss/openwiki/code-mode#open-knowledge-format">
    Emit OKF v0.1 Markdown bundles with front matter, indexes, and linked concepts.
  </Card>

  <Card title="LangSmith tracing" icon="chart-dots" href="/oss/openwiki/quickstart#trace-with-langsmith">
    Trace documentation runs with LangSmith.
  </Card>
</CardGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="player-play" href="/oss/openwiki/quickstart">
    Install OpenWiki, configure a provider, and generate your first wiki.
  </Card>

  <Card title="CLI reference" icon="terminal" href="/oss/openwiki/cli-reference">
    Review commands, flags, and connector subcommands.
  </Card>
</CardGroup>

***

<div className="source-links">
  <Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to Claude, VSCode, and more via MCP for real-time answers.
  </Callout>

  <Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/oss/openwiki/overview.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
  </Callout>
</div>