Quick Start

Quick Start

Get mcmqtt running in under 2 minutes with our zero-configuration setup. No dependencies, no complex installation, just intelligent MQTT integration that works.

Instant Execution with uvx

The fastest way to try mcmqtt is with uvx - no installation required:

# Start mcmqtt instantly
uvx mcmqtt

# Or with specific options
uvx mcmqtt --transport http --port 8080

That’s it! mcmqtt is now running and ready for MCP clients.

Add to Claude Code

For persistent use with Claude Code, add mcmqtt as an MCP server:

# Add mcmqtt to your MCP configuration
claude mcp add task-buzz "uvx mcmqtt"

# Test the connection
claude mcp test task-buzz

# Verify it's working
claude mcp list

First MQTT Connection

Once mcmqtt is running, you can immediately start using MQTT features:

Basic Pub/Sub

# In Claude Code, use the mcmqtt tools:
# 1. Connect to broker (auto-spawns if needed)
mqtt_connect({
  "host": "localhost",
  "port": 1883,
  "client_id": "my-client"
})

# 2. Subscribe to topics
mqtt_subscribe({
  "topic": "test/messages",
  "qos": 1
})

# 3. Publish messages
mqtt_publish({
  "topic": "test/messages",
  "payload": "Hello from mcmqtt!",
  "qos": 1
})

Spawn a Broker

Need a dedicated broker? mcmqtt can spawn one instantly:

# Spawn broker on custom port
spawn_mqtt_broker({
  "port": 1884,
  "config": {
    "allow_anonymous": true,
    "max_connections": 100
  }
})

Next Steps

Now that mcmqtt is running, explore these key capabilities:

🚀 Deploy Agent Swarms

Ready to coordinate AI agents? Try the fractal coordination system:

# Deploy browser testing swarm
uvx mcmqtt swarm deploy --agents browser-test --count 10

Learn about Fractal Agent Coordination →

📡 Real-time Data Streams

Build reactive applications with MQTT streams:

# Stream sensor data
mqtt_publish({
  "topic": "sensors/temperature",
  "payload": {"value": 23.5, "unit": "C", "timestamp": "2024-01-15T10:30:00Z"}
})

Build Data Streaming Applications →

🏗️ Production Deployment

Scale mcmqtt for production workloads:

# High-availability setup
uvx mcmqtt --broker-cluster 3 --persistence enabled

Deploy to Production →

Configuration Options

mcmqtt works out of the box, but you can customize behavior:

Environment Variables

export MQTT_BROKER_HOST="mqtt.example.com"
export MQTT_BROKER_PORT="1883"
export MQTT_CLIENT_ID="mcmqtt-server"
uvx mcmqtt  # Uses environment config

CLI Options

uvx mcmqtt --help  # See all options

# Common configurations:
uvx mcmqtt --transport stdio          # Default: MCP STDIO
uvx mcmqtt --transport http           # HTTP server mode
uvx mcmqtt --mqtt-host broker.local   # Connect to existing broker
uvx mcmqtt --auto-connect             # Auto-connect on startup

Troubleshooting

Connection Issues

If you can’t connect to MQTT:

# Check if broker is running
netstat -an | grep :1883

# Test with mosquitto clients
mosquitto_pub -h localhost -t test -m "hello"
mosquitto_sub -h localhost -t test

MCP Integration Issues

If Claude Code can’t find mcmqtt:

# Remove and re-add the server
claude mcp remove task-buzz
claude mcp add task-buzz "uvx mcmqtt"

# Check MCP server logs
claude mcp logs task-buzz

Performance Issues

For high-throughput scenarios:

# Increase connection limits
uvx mcmqtt --max-connections 1000 --buffer-size 65536

# Enable high-performance mode
uvx mcmqtt --performance-mode

What’s Next?

You now have mcmqtt running! Here are the most popular next steps:

1. Build Agent Networks - Create coordinated AI agent systems
2. Real-time Monitoring - Set up health checks and observability
3. Custom Tools - Extend mcmqtt with your own MCP tools
4. Production Scaling - Deploy at enterprise scale

Need Help?

• 📖 Documentation - Comprehensive guides and references
• 🐛 Issues - Report bugs or request features
• 💬 Discussions - Community support and ideas

---

Ready to revolutionize your AI coordination? mcmqtt makes it simple to build sophisticated agent networks with enterprise-grade reliability. Start with basic MQTT, scale to agent swarms. 🚀