mcpField
T
Terminator

Terminator

Remote
github.com/iris-networksDeveloper Tools10Updated 67w ago
Repository

About Terminator

Terminator is a Go MCP server for controlled terminal command execution on macOS and Linux. It serves AI integrations that need one-off commands or shell sessions that retain state, over STDIO or StreamableHTTP.

Key features

  1. 01Single-command execution with timeouts
  2. 02Persistent shell sessions
  3. 03Configurable shell selection
  4. 04STDIO and StreamableHTTP transports
  5. 05Forwarding DISPLAY for GUI commands

Use cases

  1. 01Running terminal commands from an MCP client
  2. 02Keeping directory and environment state across commands
  3. 03Launching GUI applications through commands

MCP support

Exposes execute_command, persistent_shell, and session_manager tools to MCP clients. Supports STDIO and StreamableHTTP transports.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttp://localhost:8080/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Packages
ocishanurcsenitap/terminator:latest

Tools

Details

From the project README

💻 MCP Terminal Server

Build Status Release Go Version License

🚀 A high-performance Model Context Protocol server for secure terminal execution
✨ Cross-platform • ⚡ High-performance • 🔒 Ultra-secure ✨

Terminator Terminal

Features

  • Cross-platform support: Works on macOS and Linux
  • Configurable timeouts: Set custom timeout values for command execution
  • Secure execution: Commands run in controlled environment with proper error handling
  • Platform-aware: Automatically detects and adapts to the host platform
  • Flexible shell support: Configurable shell for command execution
  • GUI application support: Automatic DISPLAY environment variable forwarding for X11 applications
  • Multiple transport modes:
    • STDIO mode for traditional MCP clients
    • 🌐 StreamableHTTP transport: Standards-compliant HTTP-based MCP transport for web integrations
  • Flexible session management: Use any session ID you want - no pre-registration required
  • 🎉 Persistent shell sessions: Maintain shell state between commands (working directory, environment variables, etc.)

Quick Start

# Build the server
make build

# Run in HTTP mode with StreamableHTTP transport
./mcp-terminal-server --http --port 8080

# Run in STDIO mode (default)
./mcp-terminal-server

🚀 Quick Testing

Test StreamableHTTP Transport
# Start server in HTTP mode
./mcp-terminal-server --http --port 8080

# Test MCP initialization
curl -X POST http://localhost:8080/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "initialize", "params": {"protocolVersion": "2024-11-05", "capabilities": {}, "clientInfo": {"name": "test-client", "version": "1.0.0"}}}'

# Test tools listing (use session ID from initialize response)
curl -X POST http://localhost:8080/mcp \
  -H "Content-Type: application/json" \
  -H "Mcp-Session-Id: <session-id>" \
  -d '{"jsonrpc": "2.0", "id": 2, "method": "tools/list"}'

# Execute a command
curl -X POST http://localhost:8080/mcp \
  -H "Content-Type: application/json" \
  -H "Mcp-Session-Id: <session-id>" \
  -d '{"jsonrpc": "2.0", "id": 3, "method": "tools/call", "params": {"name": "execute_command", "arguments": {"command": "echo Hello StreamableHTTP!"}}}'
Test Persistent Shell
# Test persistent shell functionality
chmod +x examples/persistent_shell_examples.sh
./examples/persistent_shell_examples.sh

Documentation

📚 Complete Documentation

Available Tools

  1. execute_command - Execute single commands with timeout
  2. persistent_shell - Execute commands in persistent shell sessions
  3. session_manager - Manage shell sessions (list, close)

Environment Variables

The server supports the following environment variables:

  • MCP_COMMAND_TIMEOUT - Default command timeout in seconds (default: 30)
  • MCP_SHELL - Custom shell to use for command execution (default: /bin/bash on Unix)
  • DISPLAY - X11 display for GUI applications (automatically forwarded to commands)
GUI Application Support

The server automatically forwards the DISPLAY environment variable to all executed commands, enabling GUI applications to open on the correct display. This works for both non-persistent commands and persistent shell sessions.

Example: Launch a GUI application

# Set DISPLAY when starting the server
DISPLAY=:0 ./mcp-terminal-server

# GUI applications launched through MCP will use the correct display
# e.g., xterm, firefox, gedit, etc.

Server Endpoints

When running in HTTP mode (--http flag), the server provides:

  • POST /mcp - StreamableHTTP transport endpoint for all MCP operations
    • Supports initialize, tools/list, tools/call methods
    • Requires Mcp-Session-Id header for authenticated requests
    • Returns session ID in response headers for initialize calls
MCP Protocol Support

The server implements the Model Context Protocol specification:

  • JSON-RPC 2.0 communication
  • Session management with UUID-based session IDs
  • Tool execution with structured input/output
  • Error handling with standard JSON-RPC error codes

Platform Support

  • macOS (darwin): Full support - AMD64 & ARM64
  • Linux: Full support - AMD64 & ARM64
  • Docker: Multi-architecture support via buildx

Integration

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "terminal": {
      "command": "/path/to/mcp-terminal-server",
      "env": {
        "MCP_COMMAND_TIMEOUT": "30",
        "DISPLAY": ":0"
      }
    }
  }
}
Docker
# Run AMD64 image
docker run --rm -p 8080:8080 terminal-mcp:latest-linux-amd64 --http --port 8080

# Run ARM64 image  
docker run --rm -p 8080:8080 terminal-mcp:latest-linux-arm64 --http --port 8080

# Run with persistent shell session
docker run --rm -p 8080:8080 -v /tmp:/tmp terminal-mcp:latest-linux-amd64 --http --port 8080
Other MCP Clients

The server follows the standard MCP protocol and should work with any compliant MCP client.

Development

Building
# Build for current platform
make build

# Build for all platforms
make build-all

# Build Docker images
make docker-build

# Build and push Docker images
REGISTRY=your-registry.com make docker-push

# Build DXT package for distribution
make dxt
Dependencies
  • github.com/mark3labs/mcp-go - MCP Go library
  • Go 1.23+ (automatically managed)

Production Deployment

For production environments, it's recommended to run the server with proper process management and security measures:

Process Management
  • s6: Lightweight process supervision with automatic restart
  • systemd: Full-featured service management with resource limits
  • Docker: Containerized deployment with security isolation
  • Supervisor: Python-based process management
Security Features
  • Non-root execution: Run with dedicated user account for limited privileges
  • Resource limits: CPU, memory, and file access restrictions
  • Network isolation: Bind to localhost only for enhanced security
  • AppArmor/SELinux: Additional mandatory access controls
  • File permissions: Restricted access to necessary directories only

📋 See Process Management Guide for detailed setup instructions

Integration with Iris Agent

This MCP server is used by the Iris Computer Use Agent, our advanced computer use agent for automated terminal interactions.

License

This project is provided as-is for development purposes.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.