Maestro MCP Server

Local setup required. This server has to be cloned and prepared on your machine before you register it in Claude Code.
1

Set the server up locally

Run this once to clone and prepare the server before adding it to Claude Code.

Run in terminal
cd mcp-maestro
uv venv --python 3.12 .venv
source .venv/bin/activate
uv pip install -r requirements.txt
python3 server.py
2

Register it in Claude Code

After the local setup is done, run this command to point Claude Code at the built server.

Run in terminal
claude mcp add maestro -- node "<FULL_PATH_TO_MCP_MAESTRO>/dist/index.js"

Replace <FULL_PATH_TO_MCP_MAESTRO>/dist/index.js with the actual folder you prepared in step 1.

README.md

Connects AI assistants to the Maestro research framework for multi-agent missions

Maestro MCP Server

MCP server para Maestro (AI Research Platform) con descubrimiento automático de puerto.

Puertos Oficiales

Según la documentación oficial, Maestro se ejecuta en:

  • Web UI: puerto 3000
  • Backend API: puerto 3001 (a través de nginx en 80)
  • PostgreSQL: puerto 5432

Quick Start

cd mcp-maestro
uv venv --python 3.12 .venv
source .venv/bin/activate
uv pip install -r requirements.txt
python3 server.py

Lógica de Descubrimiento

  1. Primero: Busca en puerto 3000 (web UI)
  2. Si no encuentra: Escanea puertos alternativos (8000, 10303, 8001, 5000, 8080)
  3. Si sigue sin encontrar: Pide al usuario el puerto por consola
  4. Fallback: Usa http://localhost:3000 si el usuario no responde

Configuración Manual

Si prefieres configurar manualmente, usa la variable de entorno:

export MAESTRO_BASE_URL=http://localhost:3000

O crea un archivo .env en la raíz del proyecto.

Herramientas

maestro_create_mission

  • request (string): Descripción de la misión
  • chat_id (string, optional): ID de chat
  • use_web_search (boolean): Usar búsqueda web

maestro_get_report

  • mission_id (string): ID de la misión

maestro_resume

  • mission_id (string): ID de la misión

maestro_stop

  • mission_id (string): ID de la misión

Claude Desktop

{
  "mcpServers": {
    "maestro": {
      "command": "uv",
      "args": ["--directory", "/Users/simba/Code/MCP-servers/mcp-maestro", "run", "python3", "server.py"]
    }
  }
}

MCP Server

  • Puerto: 8081
  • SSE: /sse
  • HTTP: /mcp

Tools (4)

maestro_create_missionInitiates a new research mission with a specific description.
maestro_get_reportRetrieves the final report for a completed research mission.
maestro_resumeResumes a paused or interrupted research mission.
maestro_stopStops an ongoing research mission.

Environment Variables

MAESTRO_BASE_URLThe base URL for the Maestro backend API (defaults to http://localhost:3000)

Configuration

claude_desktop_config.json
{"mcpServers": {"maestro": {"command": "uv", "args": ["--directory", "/path/to/mcp-maestro", "run", "python3", "server.py"]}}}

Try it

Create a new research mission to analyze the current state of quantum computing.
Start a research mission about renewable energy trends and include web search.
Get the final report for my mission with ID 12345.
Stop the research mission currently running with ID 67890.

Frequently Asked Questions

What are the key features of Maestro?

Orchestrates multi-agent research missions including planning and writing. Automatic discovery of Maestro service ports. Real-time research task tracking. Retrieval of structured research reports and notes.

What can I use Maestro for?

Automating deep-dive research tasks for complex topics. Managing multi-agent workflows for academic or technical writing. Integrating AI-driven research capabilities directly into Claude Desktop.

How do I install Maestro?

Install Maestro by running: cd mcp-maestro && uv venv --python 3.12 .venv && source .venv/bin/activate && uv pip install -r requirements.txt && python3 server.py

What MCP clients work with Maestro?

Maestro works with any MCP-compatible client including Claude Desktop, Claude Code, Cursor, and other editors with MCP support.

Turn this server into reusable context

Keep Maestro docs, env vars, and workflow notes in Conare so your agent carries them across sessions.

Need the old visual installer? Open Conare IDE.
Open Conare