Local Development (Python)¶

☕ Looking for Java? See Java Local Development | 📘 Looking for TypeScript? See TypeScript Local Development

Develop MCP Mesh agents locally with Python

Development Workflow¶

graph LR
    A[Install] --> B[Scaffold]
    B --> C[Run]
    C --> D[Call & Debug]
    D --> C

Guides¶

Guide	What You'll Learn
Getting Started	Install meshctl CLI and Python SDK
Scaffold Agents	Generate agents with `meshctl scaffold`
Run Agents	Start agents with hot reload, debug mode
Inspect the Mesh	View agents, tools, and dependencies
Call & Debug Tools	Call tools and trace distributed calls
Troubleshooting	Common issues and solutions

Quick Reference¶

# Install CLI
npm install -g @mcpmesh/cli

# Set up Python (at project root)
python3 -m venv .venv
source .venv/bin/activate
pip install "mcp-mesh>=0.8,<0.9"
deactivate  # meshctl auto-detects .venv

# Scaffold (interactive wizard)
meshctl scaffold

# Run
meshctl start main.py
meshctl start -w main.py          # hot reload
meshctl start --debug main.py     # debug mode

# Inspect
meshctl list                      # agents
meshctl list --tools              # all tools
meshctl status                    # wiring details

# Call
meshctl call get_weather '{"city":"London"}'
meshctl call --trace get_weather  # with tracing

# Debug
meshctl trace <trace-id>          # view call tree
meshctl logs my-agent -f          # follow logs
meshctl man --list                # built-in docs