What is docs-mcp-server?

Docs MCP Server provides an always-current, personal documentation index to combat AI hallucinations and outdated knowledge for AI coding assistants. It fetches official documentation from various sources, including websites, GitHub, and local files, allowing AI to query exact library versions.

docs-mcp-server

Active·★ 1.8k·MIT·Updated 2026-10-12

★ Trending★ Essential

Grounded Docs (docs-mcp-server) aims to reduce hallucinations and outdated answers by giving an AI coding assistant a personal, current documentation index. It fetches official documentation from websites, GitHub, npm, PyPI and local files, and the README presents it as an open-source alternative to Context7, Nia and Ref.Tools. Queries can target the exact library version in your project, and everything runs on your own machine so your code stays on your network. It can index websites, GitHub repositories, local folders and zip archives, and processes HTML, Markdown, PDF, Office documents, OpenDocument, RTF, EPUB, Jupyter notebooks and 90+ source code languages.

There are two ways to use it. The CLI-first path needs Node.js 22 or newer: `npx @arabold/docs-mcp-server@latest scrape react https://react.dev/reference/react` indexes a library, and the repository ships Agent Skills that teach assistants to use the CLI. For a long-running endpoint, run `npx @arabold/docs-mcp-server@latest`, add documentation in the web UI at `http://localhost:6280`, and point your MCP client at `http://localhost:6280/mcp` (Streamable HTTP). An embedding model is optional but recommended for semantic search, for example with `OPENAI_API_KEY`, or Ollama, Gemini or Azure instead. It is MIT licensed.

#AI Assistant#Documentation#RAG#LLM Context#Local-first#Web Scraping#Semantic Search#MCP Protocol

Features

01Fetches documentation directly from official sources on demand, ensuring current information.

02Allows querying specific library versions relevant to your project, enhancing accuracy.

03Grounds Large Language Models in real, verified documentation, minimizing incorrect outputs.

04Runs entirely on your machine, ensuring data privacy and preventing code from leaving your network.

05Indexes documentation from multiple sources (web, GitHub, local files) and supports a wide array of file formats.

Why choose it

+Version-specific queries target the exact library version in your project

+Runs locally, so your code and queries stay on your network

+Indexes websites, GitHub repos, local folders and zip archives, in many document and code formats

+CLI, Agent Skills and a Streamable HTTP MCP endpoint with a web UI

Trade-offs

!You have to scrape and index each library yourself, which can take time for large sites

!Search quality is better with an embedding model, which means an API key or a local model

!Requires Node.js 22 or newer

!The comparison with Context7, Nia and Ref.Tools is the project's own framing

Compatibility

Node.js

Runtime

Verified via docs

MCP Protocol

Integration

Verified via docs

Docker

Deployment

Verified via docs

LLM Embedding Providers

Embedding Service

Verified via docs

Quick start

$ npx @arabold/docs-mcp-server@latest

Use cases

↳Empower AI coding assistants with current, version-specific documentation to improve code generation and problem-solving.

↳Establish a private, always-current knowledge base for local LLMs, reducing dependency on external data sources.

↳Connect with MCP-compatible AI clients (e.g., Claude, Copilot) to provide them with enhanced, grounded context.

↳Scrape and index documentation from various online sources (websites, GitHub) and local files for centralized, searchable access.

Alternatives

ragflow★ 92.2k

RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs

MaxKB★ 23.0k

An open-source platform for building enterprise-grade agents. Powerful and easy to use.

headroom★ 75.2k

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. Around 20% fewer tokens for coding agents and 60-95% fewer for JSON, with the same answers. Library, proxy, and MCP server.

Scrapling★ 87.3k

🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl!

chrome-devtools-mcp★ 53.5k

Chrome DevTools for coding agents

agents-towards-production★ 21.7k

This repository delivers end-to-end, code-first tutorials covering every layer of production-grade GenAI agents, guiding you from spark to scale with proven patterns and reusable blueprints for real-world launches.

deer-flow★ 83.8k

DeerFlow is a community-driven Deep Research framework, combining language models with tools like web search, crawling, and Python execution, while contributing back to the open-source community.

Cherry Studio★ 52.6k

A powerful desktop client for multiple LLMs. Supports local and cloud models.

See all alternatives →

Featured in these lists

Best MCP Servers for Up-to-Date Docs and Code Context

Comments

No comments yet. Be the first!

docs-mcp-server

Active·★ 1.8k·MIT·Updated 2026-10-12

★ Trending★ Essential

#AI Assistant#Documentation#RAG#LLM Context#Local-first#Web Scraping#Semantic Search#MCP Protocol

Features

01Fetches documentation directly from official sources on demand, ensuring current information.

02Allows querying specific library versions relevant to your project, enhancing accuracy.

03Grounds Large Language Models in real, verified documentation, minimizing incorrect outputs.

04Runs entirely on your machine, ensuring data privacy and preventing code from leaving your network.

05Indexes documentation from multiple sources (web, GitHub, local files) and supports a wide array of file formats.

Why choose it

+Version-specific queries target the exact library version in your project

+Runs locally, so your code and queries stay on your network

+Indexes websites, GitHub repos, local folders and zip archives, in many document and code formats

+CLI, Agent Skills and a Streamable HTTP MCP endpoint with a web UI

Trade-offs

!You have to scrape and index each library yourself, which can take time for large sites

!Search quality is better with an embedding model, which means an API key or a local model

!Requires Node.js 22 or newer

!The comparison with Context7, Nia and Ref.Tools is the project's own framing

Compatibility

Node.js

Runtime

Verified via docs

MCP Protocol

Integration

Verified via docs

Docker

Deployment

Verified via docs

LLM Embedding Providers

Embedding Service

Verified via docs

Quick start

$ npx @arabold/docs-mcp-server@latest

Use cases

↳Empower AI coding assistants with current, version-specific documentation to improve code generation and problem-solving.

↳Establish a private, always-current knowledge base for local LLMs, reducing dependency on external data sources.

↳Connect with MCP-compatible AI clients (e.g., Claude, Copilot) to provide them with enhanced, grounded context.

↳Scrape and index documentation from various online sources (websites, GitHub) and local files for centralized, searchable access.

Alternatives

ragflow★ 92.2k

RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs

MaxKB★ 23.0k

An open-source platform for building enterprise-grade agents. Powerful and easy to use.

headroom★ 75.2k

Scrapling★ 87.3k

🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl!

chrome-devtools-mcp★ 53.5k

Chrome DevTools for coding agents

agents-towards-production★ 21.7k

deer-flow★ 83.8k

DeerFlow is a community-driven Deep Research framework, combining language models with tools like web search, crawling, and Python execution, while contributing back to the open-source community.