広告掲載
MCP.so
ログイン

docs-mcp-server MCP Server

@xinlei413

リポジトリ

基本情報

カテゴリ

メモリとナレッジ

ライセンス

MIT

トランスポート

stdio

作者

xinlei413

docs-mcp-server MCP Server について

概要はまだありません

docs-mcp-server MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "MCP-DOC-Server-OpenRouter": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "\\"
      ]
    }
  }
}

docs-mcp-server MCP Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is docs-mcp-server?

docs-mcp-server is an MCP server that fetches, indexes, and searches third‑party package documentation from multiple sources (websites, GitHub, npm, PyPI, local files). It processes content into chunks, supports full‑text search via SQLite FTS5, and integrates with OpenRouter API for chat/completions. Vector embeddings are currently disabled, leaving only full‑text search and chat/completions active.

How to use docs-mcp-server?

Run the server via Docker (docker run ... ghcr.io/arabold/docs-mcp-server:latest) or npx (npx -y --package=@arabold/docs-mcp-server docs-server), configured with your MCP client (e.g., Claude, Cline, Roo). An OpenAI API key is required. Use MCP tools such as scrape_docs, search_docs, and list_libraries, or use the included CLI (docs-cli).

Key features of docs-mcp-server

  • Versatile scraping: websites, GitHub, npm, PyPI, local files
  • Asynchronous job handling for scraping and indexing
  • Full‑text search with SQLite FTS5
  • OpenRouter API integration for chat/completions
  • Multiple embedding providers supported (currently disabled)
  • Simple deployment via Docker or npx

Use cases of docs-mcp-server

  • Index documentation of a library and search it via AI assistant
  • Fetch docs from multiple sources (web, GitHub, package registries)
  • Manage different versions of library documentation
  • Use chat/completions to answer questions about scraped content

よくある質問

1

Are embeddings/vector search still available?

No. The current version has removed all embedding functionality. Only full‑text search and chat/completions are provided.

2

How do I run the server?

Use Docker with the image `ghcr.io/arabold/docs-mcp-server:latest` and set the required `OPENAI_API_KEY` environment variable. Alternatively, run via npx: `npx -y --package=@arabold/docs-mcp-server docs-server`.

3

Where is the indexed data stored?

Data is stored in an SQLite database inside the container’s `/data` directory. Persist it by mounting a Docker volume (e.g., `-v docs-mcp-data:/data`). The npx mode uses a default location in your home directory.

4

What dependencies are required?

For Docker: only Docker itself. For npx: Node.js must be installed. An API key (e.g., OpenAI) is required for chat/completions and was needed for embeddings when enabled.

5

Can I use the OpenRouter API?

Yes. Set `OPENAI_API_KEY` to your OpenRouter API key and optionally `OPENAI_API_BASE` to `https://openrouter.ai/api/v1`. Many models are supported (GPT‑4.1, Claude, Gemini, etc.).

コメント

「メモリとナレッジ」の他のコンテンツ

Sooveryn

PANDAPROG

Sooveryn gives Claude Code a team of AI personas that remembers your project. Soov, the partner and team lead, keeps your decisions and how you work in a lasting memory, and finds them again by words and by meaning, from

Stele

Stele-Dev

Shared memory for AI coding agents. Claude Code, Cursor and Codex read your project's decisions, lessons and tasks before they act.

The Cloud

Teka

The Cloud is the AI-native workspace that gives every AI — Claude, ChatGPT, Gemini, and other Frontier & Free or Local models — shared memory and real agents. Not cloud computing: thecloud.so is one workspace where your

Groundwork

Rarefied-Earth

Groundwork: company memory for your AI tools. Docs public. Source closed. Public proof needs no account.

1

Codebase Memory

DeusData

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 162 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

46.7K
広告掲載