Documentation Extractor — Claude Code MCP server: features and setup
Fetches website documentation and organizes it into formats for AI agents. Use it to bring a project’s docs into an assistant’s context.
Open on GitHubWhat Documentation Extractor does
The server takes a URL, looks for llms.txt, llms-full.txt, and install.md, then splits the material into Markdown files by section. You can export the results as Markdown, JSON, YAML, or a ZIP archive. It can also handle several URLs in one batch.
To connect it, add an MCP server entry to your client configuration. The entry runs npx with the @llm-energy/mcp-server package and uses -y as an argument. The example names the server llm-energy. Extraction relies on documentation exposed by the website, including the supported llms.txt and install.md files.
What it helps with
- extract documentation from a website URL
- split documentation into Markdown files by section
- process documentation from several URLs in one batch
- export parsed docs as JSON, YAML, or ZIP
Example prompts for the agent
- Extract the docs at
docs.example.comand organize them by section. - Process documentation from several URLs and return a ZIP archive.
- Check whether this site supports
llms.txt.
How to install Documentation Extractor
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Documentation Extractor
How do I install Documentation Extractor in Claude Code?
Add an MCP server entry named llm-energy to the client configuration. Set the command to npx, include @llm-energy/mcp-server as the package, and pass -y as an argument.
What does Documentation Extractor do?
It fetches website documentation, detects files such as llms.txt and install.md, and organizes the content for AI assistants.
Which export formats does it support?
You can export documentation as Markdown, JSON, YAML, or a ZIP archive.
