convez.ai
MCP ServerDesenvolvimento WebIntermediário
AllanSJoseph

Documentation MCP Server

AllanSJoseph/documentation-mcp-server

Acesso a documentações via MCP para respostas LLM atualizadas

Ver no GitHub
InicialPluginScore · de 0 a 100

Servidor MCP em Python que fornece acesso a documentações de frameworks populares (Django, PostgreSQL, LangChain, llama-index, OpenAI) para LLMs gerarem respostas atualizadas. Integra-se com Claude Desktop, Cursor e WindSurf via ferramenta de busca que utiliza a API Serper.

Instalação

$npx convez-ai AllanSJoseph/documentation-mcp-server

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
0
Forks
0
Acompanhando
0
Saúde
28%
Último commit
há 1 ano
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Consultar documentação Django durante desenvolvimento
  • Buscar soluções PostgreSQL integradas ao chat de IA
  • Referência LangChain e llama-index em tempo real
  • Documentação OpenAI API diretamente no assistente

Veredito

Ferramenta prática para desenvolvedores que usam esses frameworks específicos. Reduz fricção ao buscar documentação durante desenvolvimento, mas está limitada a 5 frameworks e requer configuração de chaves API (Serper).

README

A tradução para o português deste README ainda está na fila.

Documentation MCP Server

A simple Python based MCP (Model Context Protocol) Server that provides documentation of some frameworks to an LLM Chatbot so that it will generate up to date results.

This is a MCP server project built with the guidance of Alejandro AO - Software & Ai 's Youtube Video Tutorial.

Refer the video here Learn MCP Servers with Python (EASY)

What this MCP Server does

  • Provides a tool named documentation
  • Uses Serper a google search API to search for specific sections of documentations.
  • Returns the documentation results from the below documentation sites.
  • Uses Beautiful Soup to parse the results

How to Setup this MCP Server

Prerequisites

  • Python 3.8 or higher
  • uv (Python package manager for MCP servers)
  • Serper API key (you can get a free trial API key for first login with 2000 credits -> Open Serper Homepage)
  • An MCP client that supports MCP servers eg: Claude Desktop, Cursor, WindSurf, etc.
  • [Optional] Node.js (for inspecting the MCP server)

Setup uv

For Windows users:

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

For Linux/Mac users (Note: This server is ran and tested on Windows):

curl -LsSf https://astral.sh/uv/install.sh | sh

Setup Serper API Key,

Create a .env file in the project directory and add your Serper API key like

Copy the api key from your Serper account and paste it in the .env file as shown below:

SERPER_API_KEY=your_serper_api_key_here
  1. Open an MCP Client (here I am using Claude Desktop).
  2. Go To Claude Settings Ctrl + ,.
  3. Go to the Developer tab.
  4. Click on Edit Config under Local MCP Servers.
  5. It highlights the claude_desktop_config.json file, Open it in your favourite text editor.
  6. Add the below code to the mcp_servers list in the JSON file save and close.
"mcpServers": {
        "documentation": {
            "command": "uv",
            "args": [
                "--directory",
                "your_project_directory_path",
                "run",
                "main.py"
            ]
        }
    }
  1. Restart Claude Desktop
  2. Open a new chat and click on tools button on the chat box and you will see the documentation tool available.
  3. Try a prompt like below:
How do I add and connect a PostgreSQL Database in Django? Use the documentation tool to get the answer.
  1. The LLM will use the documentation tool to search for the answer and return it in the chat.

Please Note that free version of Claude has a limitation of number of tokens or characters, so it may not give results as desired and can sometimes give you an error of exceeding the token limit. You can try with a paid version of Claude or use other MCP clients like Cursor, WindSurf, etc.

Refer the official MCP Documentation from Antropic for more details on how to setup MCP servers in your client For Server Developers - Model Context Protocol

Inspect this MCP Server

To inspect the tools and check if tools are working properly, run the below command in terminal on the project directory.

npx @modelcontextprotocol/inspector or uv run main.py

Thank you Alejandro AO for the amazing tutorial and guidance on building this MCP server.

Ver o original no GitHub

Bom para

Relacionados