MCP Hummingbot Server

An MCP (Model Context Protocol) server that enables Claude and Gemini CLI to interact with Hummingbot for automated cryptocurrency trading across multiple exchanges.

Installation & Configuration

Option 1: Using uv (Recommended for Development)

1. Install uv (if not already installed):

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

2. Clone and install dependencies:

   git clone https://github.com/hummingbot/mcp
   cd mcp
   uv sync
   

3. Create a .env file:

   cp .env.example .env
   

4. Edit the .env file with your Hummingbot API credentials:

   HUMMINGBOT_API_URL=http://localhost:8000
   HUMMINGBOT_USERNAME=admin
   HUMMINGBOT_PASSWORD=admin
   

5. Configure in Claude Code or Gemini CLI:

   {
     "mcpServers": {
       "hummingbot-mcp": {
         "type": "stdio",
         "command": "uv",
         "args": [
           "--directory",
           "/path/to/mcp",
           "run",
           "main.py"
         ]
       }
     }
   }
   

   Note: Make sure to replace /path/to/mcp with the actual path to your MCP directory.

Option 2: Using Docker (Recommended for Production)

1. Create a .env file:

   touch .env
   

2. Edit the .env file with your Hummingbot API credentials:

   HUMMINGBOT_API_URL=http://localhost:8000
   HUMMINGBOT_USERNAME=admin
   HUMMINGBOT_PASSWORD=admin
   

   Important: When running the MCP server in Docker and connecting to a Hummingbot API on your host:

   • Linux: Use --network host (see below) to allow the container to access localhost:8000
   • Mac/Windows: Change HUMMINGBOT_API_URL to http://host.docker.internal:8000

3. Pull the Docker image:

   docker pull hummingbot/hummingbot-mcp:latest
   

4. Configure in Claude Code or Gemini CLI:

   For Linux (using --network host):

   {
     "mcpServers": {
       "hummingbot-mcp": {
         "type": "stdio",
         "command": "docker",
         "args": [
           "run",
           "--rm",
           "-i",
           "--network",
           "host",
           "--env-file",
           "/path/to/mcp/.env",
           "-v",
           "$HOME/.hummingbot_mcp:/root/.hummingbot_mcp",
           "hummingbot/hummingbot-mcp:latest"
         ]
       }
     }
   }
   

   For Mac/Windows:

   {
     "mcpServers": {
       "hummingbot-mcp": {
         "type": "stdio",
         "command": "docker",
         "args": [
           "run",
           "--rm",
           "-i",
           "--env-file",
           "/path/to/mcp/.env",
           "-v",
           "$HOME/.hummingbot_mcp:/root/.hummingbot_mcp",
           "hummingbot/hummingbot-mcp:latest"
         ]
       }
     }
   }
   

   (Remember to set HUMMINGBOT_API_URL=http://host.docker.internal:8000 in your .env file)

   Note: Make sure to replace /path/to/mcp with the actual path to your MCP directory.

Cloud Deployment with Docker Compose

For cloud deployment where both Hummingbot API and MCP server run on the same server:

1. Create a .env file:

   touch .env
   

2. Edit the .env file with your Hummingbot API credentials:

   HUMMINGBOT_API_URL=http://localhost:8000
   HUMMINGBOT_USERNAME=admin
   HUMMINGBOT_PASSWORD=admin
   

3. Create a docker-compose.yml:

   services:
     hummingbot-api:
       container_name: hummingbot-api
       image: hummingbot/hummingbot-api:latest
       ports:
         - "8000:8000"
       volumes:
         - ./bots:/hummingbot-api/bots
         - /var/run/docker.sock:/var/run/docker.sock
       environment:
         - USERNAME=admin
         - PASSWORD=admin
         - BROKER_HOST=emqx
         - DATABASE_URL=postgresql+asyncpg://hbot:hummingbot-api@postgres:5432/hummingbot_api
       networks:
         - emqx-bridge
       depends_on:
         - postgres
   
     mcp-server:
       container_name: hummingbot-mcp
       image: hummingbot/hummingbot-mcp:latest
       stdin_open: true
       tty: true
       env_file:
         - .env
       environment:
         - HUMMINGBOT_API_URL=http://hummingbot-api:8000
       depends_on:
         - hummingbot-api
       networks:
         - emqx-bridge
   
     # Include other services from hummingbot-api docker-compose.yml as needed
     emqx:
       container_name: hummingbot-broker
       image: emqx:5
       restart: unless-stopped
       environment:
         - EMQX_NAME=emqx
         - EMQX_HOST=node1.emqx.local
         - EMQX_CLUSTER__DISCOVERY_STRATEGY=static
         - EMQX_CLUSTER__STATIC__SEEDS=[emqx@node1.emqx.local]
         - EMQX_LOADED_PLUGINS="emqx_recon,emqx_retainer,emqx_management,emqx_dashboard"
       volumes:
         - emqx-data:/opt/emqx/data
         - emqx-log:/opt/emqx/log
         - emqx-etc:/opt/emqx/etc
       ports:
         - "1883:1883"
         - "8883:8883"
         - "8083:8083"
         - "8084:8084"
         - "8081:8081"
         - "18083:18083"
         - "61613:61613"
       networks:
         emqx-bridge:
           aliases:
             - node1.emqx.local
       healthcheck:
         test: [ "CMD", "/opt/emqx/bin/emqx_ctl", "status" ]
         interval: 5s
         timeout: 25s
         retries: 5
   
     postgres:
       container_name: hummingbot-postgres
       image: postgres:15
       restart: unless-stopped
       environment:
         - POSTGRES_DB=hummingbot_api
         - POSTGRES_USER=hbot
         - POSTGRES_PASSWORD=hummingbot-api
       volumes:
         - postgres-data:/var/lib/postgresql/data
       ports:
         - "5432:5432"
       networks:
         - emqx-bridge
       healthcheck:
         test: ["CMD-SHELL", "pg_isready -U hbot -d hummingbot_api"]
         interval: 10s
         timeout: 5s
         retries: 5
   
   networks:
     emqx-bridge:
       driver: bridge
   
   volumes:
     emqx-data: { }
     emqx-log: { }
     emqx-etc: { }
     postgres-data: { }
   

4. Deploy:

   docker compose up -d
   

5. Configure in Claude Code or Gemini CLI to connect to existing container:

   {
     "mcpServers": {
       "hummingbot-mcp": {
         "type": "stdio",
         "command": "docker",
         "args": [
           "exec",
           "-i",
           "hummingbot-mcp",
           "uv",
           "run",
           "main.py"
         ]
       }
     }
   }
   

   Note: Replace hummingbot-mcp with your actual container name. You can find the container name by running:

   docker ps
   

Server Configuration

On first run, the server creates a default configuration from environment variables (or uses http://localhost:8000 with default credentials). Configuration is stored in ~/.hummingbot_mcp/server.yml.

Using the configure_server Tool

# Show the current server configuration
configure_server()

# Update the host and port
configure_server(host="192.168.1.100", port=8001)

# Update credentials
configure_server(username="admin", password="secure_password")

# Update everything at once
configure_server(
    name="production",
    host="prod-server",
    port=8000,
    username="admin",
    password="secure_password"
)

Only the provided parameters are changed; omitted ones keep their current values. The client automatically reconnects after any update.

Environment Variables

The following environment variables can be set in your .env file for the MCP server:

Variable	Default	Description
HUMMINGBOT_API_URL	http://localhost:8000	Initial default API server URL (used only on first run)
HUMMINGBOT_USERNAME	admin	Initial username (used only on first run)
HUMMINGBOT_PASSWORD	admin	Initial password (used only on first run)
HUMMINGBOT_TIMEOUT	30.0	Connection timeout in seconds
HUMMINGBOT_MAX_RETRIES	3	Maximum number of retry attempts
HUMMINGBOT_RETRY_DELAY	2.0	Delay between retries in seconds
HUMMINGBOT_LOG_LEVEL	INFO	Logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)

Note: After initial setup, use the configure_server tool to update the server connection. Environment variables are only used to create the initial default configuration.

Requirements

• Python 3.11+
• Running Hummingbot API server
• Valid Hummingbot API credentials

Available Tools

The MCP server provides tools for:

Server Management

• configure_server: View or update the active Hummingbot API server connection
  • No parameters: show current server config
  • Any parameters: update and reconnect
  • Configuration persists in ~/.hummingbot_mcp/server.yml

Trading & Account Management

• Account management and connector setup
• Portfolio balances and distribution
• Order placement and management
• Position management
• Market data (prices, order books, candles)
• Funding rates
• Bot deployment and management
• Controller configuration

Development

To run the server in development mode:

uv run main.py

To run tests:

uv run pytest

Troubleshooting

The MCP server now provides comprehensive error messages to help diagnose connection and authentication issues:

Connection Errors

If you see error messages like:

• ❌ Cannot reach Hummingbot API at <url> - The API server is not running or not accessible
• ❌ Authentication failed when connecting to Hummingbot API - Incorrect username or password
• ❌ Failed to connect to Hummingbot API - Generic connection failure

The error messages will include:

• The exact URL being used
• Your configured username (password is masked)
• Specific suggestions on how to fix the issue
• References to tools like configure_server

Common Solutions

1. API Not Running:

   • Ensure your Hummingbot API server is running
   • Verify the API is accessible at the configured URL

2. Wrong Credentials:

   • Use configure_server tool to update server credentials
   • Or check your .env file configuration

3. Wrong URL:

   • Use configure_server tool to update the server URL
   • For Docker on Mac/Windows, use host.docker.internal instead of localhost

4. Docker Network Issues:

   • On Linux, use --network host in your Docker configuration
   • On Mac/Windows, use host.docker.internal:8000 as the API URL

Error Prevention

The MCP server will:

• Not retry on authentication failures (401 errors) - it will immediately tell you the credentials are wrong
• Retry on connection failures with helpful messages about what might be wrong
• Provide context about whether you're running in Docker and suggest appropriate fixes
• Guide you to the right tools (configure_server) to fix issues