Skip to content

CLI Reference

Complete command-line interface documentation for the MCP Platform Templates system.

The MCP Template CLI provides a comprehensive set of commands for deploying, managing, and working with Model Context Protocol (MCP) servers. It supports multiple deployment backends, rich configuration options, and advanced features like tool discovery and integration examples.

Quick Start

# Install MCP Templates
pip install mcp-templates

# List available templates
mcpt list

# Deploy a template with defaults
mcpt deploy demo

# View template configuration options
mcpt config demo

# Discover tools from template
mcpt> tools demo

Global Options

Option Description
--help, -h Show help message and exit
--version Show version information
--verbose, -v Enable verbose output with detailed logging
--quiet, -q Suppress output except errors

Command Reference

Template Management

  • create - Create new MCP server templates with interactive guidance
  • deploy - Deploy MCP server templates to infrastructure
  • list - List available templates and deployments
  • stop - Stop deployed templates or specific deployments

Tool Discovery & Interaction

  • connect - Connect to deployed MCP servers with LLM clients

Configuration & Monitoring

  • config - View and validate configuration settings
  • logs - View deployment logs and monitor status
  • logs - Access deployment logs and monitoring

Configuration System

The CLI supports comprehensive template creation alongside deployment and configuration management:

Template Creation & Development

# Interactive template creation
mcpt create

# Create from existing image
mcpt create --from-image mcp/filesystem my-filesystem

# Non-interactive creation
mcpt create --config-file template-config.json --non-interactive

Configuration Precedence

The CLI supports multiple configuration methods with clear precedence:

Environment Variables > CLI Options > Config File > Template Defaults

Configuration Sources

  1. Environment Variables: MCP_* prefixed variables
  2. CLI Options: --config key=value and --env KEY=VALUE
  3. Configuration Files: JSON/YAML files via --config-file
  4. Template Defaults: Built-in template configurations

Double-Underscore Notation

Use double underscores for nested configuration:

# Standard format
mcpt deploy filesystem --config read_only_mode=true

# Nested format (equivalent)
mcpt deploy filesystem --config security__read_only=true

# Template-prefixed format
mcpt deploy filesystem --config filesystem__security__read_only=true

Common Usage Patterns

Development Workflow

# Create and test new template
mcpt create my-server
mcpt deploy my-server --config debug=true
mcpt logs my-server --follow

# Test tools
mcpt> tools my-server
mcpt connect my-server --llm claude

Production Deployment

# Deploy with production config
mcpt deploy filesystem \
  --config-file production.json \
  --name prod-filesystem \
  --no-pull

# Monitor deployment
mcpt status prod-filesystem
mcpt logs prod-filesystem --tail 100

Tool Discovery

# Unified tool discovery command
mcpt> tools demo                              # From templates
mcpt> tools --image mcp/filesystem /tmp       # From Docker images

# Get integration examples
mcpt connect demo --llm vscode

Environment Variables

Variable Description Default
MCP_TEMPLATE_DEFAULT_BACKEND Default deployment backend docker
MCP_TEMPLATE_CONFIG_PATH Default configuration directory ~/.mcp-template
MCP_DEBUG Enable debug logging false
DOCKER_HOST Docker daemon host (for Docker backend) unix:///var/run/docker.sock

Exit Codes

Code Meaning
0 Success
1 General error
2 Invalid argument or configuration
3 Template not found
4 Deployment error
5 Backend not available

Error Handling

The CLI provides comprehensive error handling with helpful messages:

# Template not found
❌ Template 'nonexistent' not found
Available templates: demo, filesystem, postgres-server

# Configuration error
❌ Invalid configuration: security.max_file_size must be a number
   Given: "unlimited"
   Expected: integer

# Backend error
❌ Docker daemon not available
   Please ensure Docker is installed and running

Integration Examples

Claude Desktop

{
  "mcpServers": {
    "my-server": {
      "command": "docker",
      "args": ["exec", "-i", "CONTAINER_NAME", "python", "-m", "src.server"]
    }
  }
}

VS Code

Add to .vscode/settings.json:

{
  "mcp.servers": {
    "my-server": {
      "command": "docker",
      "args": ["exec", "-i", "CONTAINER_NAME", "python", "-m", "src.server"]
    }
  }
}

Python

import asyncio
from mcp_template.client import MCPClient

async def main():
    client = MCPClient("container-name")
    await client.connect()

    tools = await client.list_tools()
    result = await client.call_tool("example_tool", {"param": "value"})

    await client.disconnect()

asyncio.run(main())

Support & Troubleshooting

For common issues and troubleshooting: - Check Troubleshooting Guide - Review FAQ - Visit GitHub Issues

Next Steps