LangChain 开源了 OpenWiki——一个专门给 agent 用的 CLI，自动为你的代码库写文档并持续维护。不是让你手写文档，是让 agent 写、agent 维护、agent 参考。

OpenWiki：LangChain 开源的 Agent 自维护文档系统

LangChain 昨天发了个新项目，477 star，MIT 开源。OpenWiki 是一个 CLI，专门为 agent 设计——自动给代码库生成文档并持续更新。

安装： npm install -g openwiki

怎么用： openwiki –init 配置模型和 API key → openwiki 生成文档到 openwiki/ 目录 → 之后自动在 AGENTS.md / CLAUDE.md 里追加提示，让 coding agent 知道参考这个 wiki

关键设计： • 支持 GLM 5.2、Kimi K2.6、Sonnet 5 等，默认预配好的模型列表 • 支持 OpenRouter、Fireworks、Baseten、OpenAI、Anthropic • GitHub Action 集成——每天自动开 PR 更新文档 • 文档目录存在就做增量更新，不存在就从头创建

跟我们今天翻译的主题全部呼应： • Karpathy 的 LLM Knowledge Base（同样的概念，这次做成了自动化工具） • EXM7777 的 Agent OS 原则“keep memory in files“ • 20 个 Loop 模式里的 Memory Update Loop

这个工具把“让 AI 记住你的代码库“这件事从概念推进到了 npm install 一步。LangChain 说它 built specifically for agents——不是为人类写的 wiki。477 star 说明这件事正好是痛点。

原文：https://github.com/langchain-ai/openwiki…

#OpenWiki #LangChain #Agent文档

langchain-ai/openwiki

Source: https://github.com/langchain-ai/openwiki

OpenWiki

OpenWiki is a CLI that writes and maintains documentation for your codebase, built specifically for agents.

Install

npm install -g openwiki

Quick Start

Initialize OpenWiki, configure your model and API key, then generate documentation

openwiki --init

Then to ensure your documentation stays up-to-date, add the GitHub action to your repository to automatically open a PR once a day with documentation updates: openwiki-update.yml

Copy the contents of that file into .github/workflows/openwiki-update.yml in your repository.

Usage

Start the interactive CLI:

openwiki

Start OpenWiki with an initial request:

openwiki "Please generate documentation for this repository"

Run a single command and exit:

openwiki -p "Summarize what you can do"

Initialize OpenWiki:

openwiki --init

Update existing documentation:

openwiki --update

Show help:

openwiki --help

openwiki creates initial documentation in openwiki/ when no wiki exists. If openwiki/ already exists, it refreshes that documentation from repository changes. By default, the CLI stays open after each run so you can send follow-up messages. Use -p or --print for a one-shot non-interactive run that prints the final assistant output.

openwiki will automatically append prompting to your AGENTS.md and/or CLAUDE.md files to instruct your coding agent to reference it when searching for context. If the file does not already exist in your repository, OpenWiki will create it for you.

On the first interactive run, OpenWiki will have you configure your inference provider, API key, and LLM. You will also be able to set a LangSmith API key to trace your OpenWiki runs to a LangSmith tracing project named “openwiki” (optional).

These configuration options and secrets will be saved to ~/.openwiki/.env on your local machine.

Customizing

OpenWiki supports OpenRouter, Fireworks, Baseten, OpenAI and Anthropic out of the box. By default, there are a few models pre-defined (GLM 5.2, Kimi K2.6, Sonnet 5, etc) but for each inference provider, OpenWiki will allow you to specify your own custom model ID.

If there’s an inference provider or model you’d like to see added, please open a PR!