How do I use mcp-server-python?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-server-python search knowledge base for pricing" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

mcp-server-python

by Dimitrearaujo

Overview Schema Related Servers Score Discussions

Python

Local

mcp-server-python

MCP server em Python com tools customizadas para agentes IA — knowledge base local, contexto de negócio e sumarização de texto.

O que é MCP?

MCP (Model Context Protocol) é um protocolo aberto criado pela Anthropic que permite que agentes IA (como o Claude) se conectem a servidores externos para acessar dados, executar ações e usar ferramentas customizadas.

Com um MCP server, você pode:

Dar ao Claude acesso à sua base de conhecimento interna
Expor dados de negócio sem precisar colocá-los no prompt manualmente
Criar pipelines de processamento de texto reutilizáveis
Integrar qualquer sistema externo como uma "tool" para o agente

Este servidor expõe 3 tools prontas:

Tool	O que faz
`search_knowledge_base`	Busca documentos relevantes numa SQLite local via cosine similarity
`summarize_text`	Formata um prompt de sumarização (sem chamar API — retorna o prompt pronto)
`get_business_context`	Retorna contexto de negócio de um JSON local

Related MCP server: mcplens

Instalação

Pré-requisitos

Python 3.12+
pip

Passos

# Clone o repositório
git clone https://github.com/dimitrearaujo/mcp-server-python.git
cd mcp-server-python

# Crie e ative o ambiente virtual
python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# ou
.venv\Scripts\activate  # Windows

# Instale as dependências
pip install -r requirements.txt

# Configure as variáveis de ambiente
cp .env.example .env
# Edite o .env conforme necessário

Configuração

Edite o arquivo .env:

BUSINESS_CONTEXT_PATH=./business_context.json
KB_DATABASE_PATH=./data/knowledge_base.db
KB_MAX_RESULTS=5

Edite o business_context.json com os dados da sua empresa.

Como rodar

python server.py

O servidor inicia via stdio e aguarda chamadas de tools. A saída de log vai para stderr.

Integração com Claude Desktop

Adicione ao seu claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-server-python": {
      "command": "python",
      "args": ["/caminho/para/mcp-server-python/server.py"],
      "env": {
        "BUSINESS_CONTEXT_PATH": "/caminho/para/mcp-server-python/business_context.json",
        "KB_DATABASE_PATH": "/caminho/para/mcp-server-python/data/knowledge_base.db"
      }
    }
  }
}

Localização do arquivo de configuração:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json

Após reiniciar o Claude Desktop, as tools aparecerão disponíveis no chat.

Populando a Knowledge Base

Use o Python interativo ou um script para inserir documentos:

from mcp_server.knowledge_base import KnowledgeBase

kb = KnowledgeBase(db_path="./data/knowledge_base.db")

# Inserir documentos
kb.insert(
    title="Política de Atendimento",
    content="Atendemos de segunda a sexta, das 8h às 18h. Urgências via WhatsApp.",
    category="atendimento"
)

kb.insert(
    title="Tabela de Preços",
    content="Consulta simples: R$ 150. Consulta especializada: R$ 250. Retorno em 30 dias: gratuito.",
    category="precos"
)

# Buscar
results = kb.search("qual o preço da consulta?")
for r in results:
    print(f"[{r['score']:.2f}] {r['title']}: {r['content'][:80]}...")

Como adicionar novas tools

Defina a tool em mcp_server/tools.py, adicionando um item a TOOL_DEFINITIONS:

{
    "name": "minha_nova_tool",
    "description": "Descrição clara do que a tool faz.",
    "input_schema": {
        "type": "object",
        "properties": {
            "parametro": {"type": "string", "description": "..."},
        },
        "required": ["parametro"],
    },
}

Implemente o handler assíncrono:

async def handle_minha_nova_tool(arguments: Dict[str, Any]) -> List[TextContent]:
    resultado = faz_algo(arguments["parametro"])
    return [TextContent(type="text", text=json.dumps(resultado, ensure_ascii=False))]

Registre o handler no dicionário TOOL_HANDLERS:

TOOL_HANDLERS = {
    # ... tools existentes ...
    "minha_nova_tool": handle_minha_nova_tool,
}

Pronto. A tool já aparece automaticamente via list_tools.

Estrutura do projeto

mcp-server-python/
├── .env.example              # Variáveis de ambiente necessárias
├── .gitignore
├── .github/workflows/ci.yml  # CI: syntax check + unit tests
├── README.md
├── requirements.txt
├── server.py                 # Entry point — inicia o MCP server via stdio
├── mcp_server/
│   ├── __init__.py
│   ├── tools.py              # Definição e registro das 3 tools
│   ├── knowledge_base.py     # SQLite + busca vetorial TF cosine
│   ├── context.py            # Carrega business_context.json
│   └── prompts.py            # Templates de prompt
├── business_context.json     # Contexto de negócio (edite com seus dados)
└── data/                     # Pasta para arquivos da KB (SQLite)
    └── .gitkeep

Tecnologias

MCP SDK — protocolo de comunicação com agentes IA
SQLite — armazenamento local da knowledge base (sem servidor externo)
TF Cosine Similarity — busca vetorial simples sem embeddings externos
python-dotenv — gerenciamento de variáveis de ambiente

Desenvolvido por

CD Tech — Automação e Agentes IA para Pequenos Negócios Fortaleza, CE — Brasil cd-tech-lp.pages.dev

mcp-server-python

Python MCP server with custom tools for AI agents — local knowledge base, business context and text summarization.

What is MCP?

MCP (Model Context Protocol) is an open protocol created by Anthropic that lets AI agents (like Claude) connect to external servers to access data, execute actions and use custom tools.

With an MCP server, you can:

Give Claude access to your internal knowledge base
Expose business data without having to paste it into the prompt manually
Build reusable text-processing pipelines
Integrate any external system as a "tool" for the agent

This server exposes 3 ready-to-use tools:

Tool	What it does
`search_knowledge_base`	Searches relevant documents in a local SQLite via cosine similarity
`summarize_text`	Formats a summarization prompt (no API call — returns the ready prompt)
`get_business_context`	Returns business context from a local JSON file

Installation

Prerequisites

Python 3.12+
pip

Steps

# Clone the repository
git clone https://github.com/dimitrearaujo/mcp-server-python.git
cd mcp-server-python

# Create and activate the virtual environment
python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# or
.venv\Scripts\activate  # Windows

# Install dependencies
pip install -r requirements.txt

# Configure environment variables
cp .env.example .env
# Edit .env as needed

Configuration

Edit the .env file:

BUSINESS_CONTEXT_PATH=./business_context.json
KB_DATABASE_PATH=./data/knowledge_base.db
KB_MAX_RESULTS=5

Edit business_context.json with your company's data.

How to run

python server.py

The server starts via stdio and waits for tool calls. Log output goes to stderr.

Claude Desktop integration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-server-python": {
      "command": "python",
      "args": ["/path/to/mcp-server-python/server.py"],
      "env": {
        "BUSINESS_CONTEXT_PATH": "/path/to/mcp-server-python/business_context.json",
        "KB_DATABASE_PATH": "/path/to/mcp-server-python/data/knowledge_base.db"
      }
    }
  }
}

Config file location:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json

After restarting Claude Desktop, the tools will show up as available in the chat.

Populating the Knowledge Base

Use interactive Python or a script to insert documents:

from mcp_server.knowledge_base import KnowledgeBase

kb = KnowledgeBase(db_path="./data/knowledge_base.db")

# Insert documents
kb.insert(
    title="Support Policy",
    content="We're open Monday to Friday, 8am to 6pm. Urgent requests via WhatsApp.",
    category="support"
)

kb.insert(
    title="Price Table",
    content="Basic consultation: $30. Specialized consultation: $50. Follow-up within 30 days: free.",
    category="pricing"
)

# Search
results = kb.search("what's the price of a consultation?")
for r in results:
    print(f"[{r['score']:.2f}] {r['title']}: {r['content'][:80]}...")

How to add new tools

Define the tool in mcp_server/tools.py, adding an entry to TOOL_DEFINITIONS:

{
    "name": "my_new_tool",
    "description": "Clear description of what the tool does.",
    "input_schema": {
        "type": "object",
        "properties": {
            "parameter": {"type": "string", "description": "..."},
        },
        "required": ["parameter"],
    },
}

Implement the async handler:

async def handle_my_new_tool(arguments: Dict[str, Any]) -> List[TextContent]:
    result = do_something(arguments["parameter"])
    return [TextContent(type="text", text=json.dumps(result, ensure_ascii=False))]

Register the handler in the TOOL_HANDLERS dict:

TOOL_HANDLERS = {
    # ... existing tools ...
    "my_new_tool": handle_my_new_tool,
}

Done. The tool automatically shows up via list_tools.

Project structure

mcp-server-python/
├── .env.example              # Required environment variables
├── .gitignore
├── .github/workflows/ci.yml  # CI: syntax check + unit tests
├── README.md
├── requirements.txt
├── server.py                 # Entry point — starts the MCP server via stdio
├── mcp_server/
│   ├── __init__.py
│   ├── tools.py              # Definition and registration of the 3 tools
│   ├── knowledge_base.py     # SQLite + TF cosine vector search
│   ├── context.py            # Loads business_context.json
│   └── prompts.py            # Prompt templates
├── business_context.json     # Business context (edit with your data)
└── data/                     # Folder for KB files (SQLite)
    └── .gitkeep

Technologies

MCP SDK — communication protocol with AI agents
SQLite — local knowledge base storage (no external server)
TF Cosine Similarity — simple vector search without external embeddings
python-dotenv — environment variable management

Developed by

CD Tech — AI Automation & Agents for Small Businesses Fortaleza, Brazil cd-tech-lp.pages.dev

← Back to profile

This server cannot be installed

license - not found

quality - not tested

maintenance

How are these scores calculated?

Maintenance

–Maintainers

–Response time

–Release cycle

–Releases (12mo)

Commit activity

Resources

GitHub Repository

Need Help?

Related Servers

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

Your AI Chatbot Just Exposed Your CEO's Salary to an Intern
By Om-Shree-0709 on July 2, 2026.
Agent Identity
MCP Security
OAuth Delegation
Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)
By Om-Shree-0709 on June 30, 2026.
Agentic Ai
Prompt Injection
WebAssembly
Lightport: Open-Sourcing Glama's AI Gateway
By punkpeye on April 27, 2026.
OpenAI
open source

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Dimitrearaujo/mcp-server-python'

If you have feedback or need assistance with the MCP directory API, please join our Discord server