PluginBench
MCP Server
Active
MIT

io.github.mkpvishnu/terminal-mcp MCP Server

io.github.mkpvishnu/terminal-mcp

Give your AI a real terminal with persistent sessions, interactive REPLs, SSH, and TUI app support.

What is the io.github.mkpvishnu/terminal-mcp MCP server?

The terminal-mcp MCP server provides AI agents with real, persistent terminal sessions backed by PTYs. It enables interactive command execution, SSH connections, REPL interactions, database CLI sessions, and TUI app navigation—capabilities unavailable in isolated subprocess tools. Works with Claude, VS Code, Cursor, and other MCP-compatible clients.

terminal-mcp bridges the gap between AI coding tools and real terminal workflows. Instead of isolated one-off commands, it maintains persistent terminal sessions where state carries over between calls. This unlocks SSH remote work, interactive Python/Node/Ruby REPLs, database CLI queries, TUI app navigation (htop, vim, fzf), long-running process monitoring, and more—exactly like a human at a terminal.

How to install io.github.mkpvishnu/terminal-mcp

Copy-paste configuration for popular MCP clients.

transport: stdio
Config generated by PluginBench — verify against the source before use.
Environment / auth
  • TERMINAL_MCP_MAX_SESSIONS

    Maximum concurrent sessions

  • TERMINAL_MCP_IDLE_TIMEOUT

    Seconds before auto-close (default 1800)

~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "terminal-mcp": {
      "command": "uvx",
      "args": [
        "terminal-mcp"
      ],
      "env": {
        "TERMINAL_MCP_MAX_SESSIONS": "<YOUR_TERMINAL_MCP_MAX_SESSIONS>",
        "TERMINAL_MCP_IDLE_TIMEOUT": "<YOUR_TERMINAL_MCP_IDLE_TIMEOUT>"
      }
    }
  }
}

Tools & capabilities

Tools this server exposes to the agent.

  • session_create — Spawn a persistent terminal session with a specified command
  • session_send — Send text, keys, or control characters to an active session
  • session_read — Read output from a session in stream, snapshot, auto, or diff modes
  • session_interact — Send input and read output in a single call to reduce round trips
  • session_wait_for — Wait for a regex pattern to appear in session output with timeout
  • session_exec — Execute a one-shot command without creating a persistent session
  • session_close — Close a session gracefully
  • session_resize — Resize terminal dimensions dynamically
  • session_list — List all active sessions

Use cases

  • SSH into remote servers and run multiple commands in sequence with persistent connection state
  • Run interactive Python, Node, or Ruby REPLs and execute code across multiple statements
  • Query databases interactively using psql, mysql, redis-cli, or other database CLIs
  • Navigate and control TUI applications like htop, vim, fzf with arrow keys and special keys
  • Monitor long-running builds or processes and wait for completion patterns

io.github.mkpvishnu/terminal-mcp MCP server FAQ

What is terminal-mcp?

terminal-mcp is an MCP server that gives AI agents real, persistent terminal sessions with PTY support. Unlike isolated subprocess tools, it maintains state across commands, enabling SSH, REPLs, database CLIs, TUI apps, and long-running processes.

Is terminal-mcp free?

Yes, terminal-mcp is open-source under the MIT license and free to use.

How do I install it in Cursor?

Click the one-click install badge in the README, or manually add to `.vscode/mcp.json` with command `uvx` and args `["terminal-mcp"]`.

How do I install it in Claude Desktop?

Add to your `claude_desktop_config.json` file under `mcpServers` with command `uvx` and args `["terminal-mcp"]`, then restart Claude Desktop.

Do I need authentication or API keys?

No, terminal-mcp requires no authentication or API keys. It runs locally and manages terminal sessions on your machine.

What platforms does it support?

terminal-mcp supports Linux, macOS, and Windows, and works with Claude Code, Claude Desktop, VS Code, Cursor, and Windsurf.

README (reference)

Source of truth, from the repository.

<!-- mcp-name: io.github.mkpvishnu/terminal-mcp --> <p align="center"> <img src="assets/banner.svg" width="800" alt="terminal-mcp banner"/> </p> <h3 align="center">Give your AI a real terminal. Persistent sessions. Interactive programs. Zero limitations.</h3> <p align="center"> <a href="https://pypi.org/project/terminal-mcp/"><img src="https://img.shields.io/pypi/v/terminal-mcp.svg" alt="PyPI"/></a> <a href="https://www.python.org/downloads/"><img src="https://img.shields.io/badge/python-3.10%2B-blue.svg" alt="Python 3.10+"/></a> <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-green.svg" alt="License: MIT"/></a> <a href="https://github.com/mkpvishnu/terminal-mcp/actions/workflows/ci.yml"><img src="https://github.com/mkpvishnu/terminal-mcp/actions/workflows/ci.yml/badge.svg" alt="CI"/></a> <a href="https://github.com/mkpvishnu/terminal-mcp/actions/workflows/codeql.yml"><img src="https://github.com/mkpvishnu/terminal-mcp/actions/workflows/codeql.yml/badge.svg" alt="CodeQL"/></a> </p> <p align="center"> <a href="https://insiders.vscode.dev/redirect/mcp/install?name=terminal-mcp&config=%7B%22command%22%3A%22uvx%22%2C%22args%22%3A%5B%22terminal-mcp%22%5D%7D"><img src="https://img.shields.io/badge/VS_Code-Install-007ACC?logo=visual-studio-code&logoColor=white" alt="Install in VS Code"/></a> <a href="https://insiders.vscode.dev/redirect/mcp/install?name=terminal-mcp&config=%7B%22command%22%3A%22uvx%22%2C%22args%22%3A%5B%22terminal-mcp%22%5D%7D"><img src="https://img.shields.io/badge/VS_Code_Insiders-Install-24bfa5?logo=visual-studio-code&logoColor=white" alt="Install in VS Code Insiders"/></a> <a href="cursor://anysphere.cursor-mcp/install?name=terminal-mcp&config=eyJjb21tYW5kIjoidXZ4IiwiYXJncyI6WyJ0ZXJtaW5hbC1tY3AiXX0="><img src="https://img.shields.io/badge/Cursor-Install-F37626?logo=cursor&logoColor=white" alt="Install in Cursor"/></a> <a href="#install-in-claude-desktop"><img src="https://img.shields.io/badge/Claude_Desktop-Install-cc785c?logo=claude&logoColor=white" alt="Install in Claude Desktop"/></a> </p> <p align="center"> <img src="assets/demo.gif" alt="terminal-mcp demo" width="700"/> </p>

The Problem

Every AI coding tool hits the same wall: no real terminal access.

Claude Code's Bash tool, GitHub Copilot, and Codex all run commands in isolated subprocesses. Each command starts fresh. No state carries over. That means:

  • No SSH sessions - Can't connect to a remote server and run multiple commands
  • No REPLs - Can't use Python, Node, or Ruby interpreters interactively
  • No database CLIs - Can't maintain a psql, mysql, or redis-cli connection
  • No TUI apps - Can't navigate htop, vim, or fzf with arrow keys
  • No long-running processes - Can't monitor builds, watch logs, or run dev servers

The Solution

terminal-mcp gives AI agents a real terminal. Persistent PTY sessions that survive across tool calls. Send commands, read output, press keys, navigate TUIs - exactly like a human at a terminal.

uvx terminal-mcp

One command. Works with Claude Code, Claude Desktop, VS Code, Cursor, and Windsurf.


Quick Start

1. Install (30 seconds)

# No install needed - run directly
uvx terminal-mcp

# Or install globally
pip install terminal-mcp

2. Connect to Your AI Client

<details open> <summary><strong>Claude Code</strong></summary>

Add to ~/.claude.json or project .mcp.json:

{
  "mcpServers": {
    "terminal": {
      "command": "uvx",
      "args": ["terminal-mcp"]
    }
  }
}
</details> <details> <summary><strong>Claude Desktop</strong></summary>

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "terminal": {
      "command": "uvx",
      "args": ["terminal-mcp"]
    }
  }
}
</details> <details> <summary><strong>VS Code / Cursor</strong></summary>

Click the one-click install badge above, or add to .vscode/mcp.json:

{
  "servers": {
    "terminal-mcp": {
      "command": "uvx",
      "args": ["terminal-mcp"]
    }
  }
}
</details> <details> <summary><strong>Windsurf</strong></summary>

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "terminal": {
      "command": "uvx",
      "args": ["terminal-mcp"]
    }
  }
}
</details>

3. Verify

session_exec  exec="echo hello from terminal-mcp"

What Can You Do With It?

SSH Into Remote Servers

session_create   command="ssh user@prod-server.com"   label="prod"
session_interact session_id="a1b2c3d4"  input="df -h"  wait_for="\$"
session_interact session_id="a1b2c3d4"  input="docker ps"  wait_for="\$"
session_close    session_id="a1b2c3d4"

Run Interactive REPLs

session_create   command="python3"  label="python"
session_interact session_id="e5f6g7h8"  input="import pandas as pd"  wait_for=">>>"
session_interact session_id="e5f6g7h8"  input="df = pd.read_csv('data.csv')"  wait_for=">>>"
session_interact session_id="e5f6g7h8"  input="df.describe()"  wait_for=">>>"
session_close    session_id="e5f6g7h8"

Query Databases

session_create   command="psql -U admin mydb"  label="db"
session_interact session_id="x1y2z3w4"  input="SELECT count(*) FROM users;"  wait_for="row"
session_interact session_id="x1y2z3w4"  input="\dt"  wait_for="#"
session_close    session_id="x1y2z3w4"

Navigate TUI Apps

session_create   command="htop"  label="monitor"
session_read     session_id="a1b2c3d4"
# Auto-detects TUI, returns screen snapshot

session_send     session_id="a1b2c3d4"  key="F6"
session_read     session_id="a1b2c3d4"  mode="diff"
# Returns only changed lines - saves tokens

session_send     session_id="a1b2c3d4"  key="F10"
session_close    session_id="a1b2c3d4"

Monitor Long-Running Builds

session_create   command="bash"  label="build"
session_send     session_id="a1b2c3d4"  input="npm run build"
session_wait_for session_id="a1b2c3d4"  pattern="Build complete|ERROR"  timeout=120

Run One-Off Commands

session_exec  exec="git log --oneline -10"
session_exec  exec="docker compose ps"  timeout=10

Features at a Glance

FeatureWhat It Does
Persistent SessionsReal PTY sessions that survive across tool calls
Send + Read in One Callsession_interact halves LLM round trips
Pattern-Based Readswait_for blocks until regex matches - no guessing timeouts
Auto TUI DetectionDetects htop, vim, etc. and auto-switches to screen snapshot mode
Output Diff ModeReturns only changed screen lines - minimizes tokens
Special KeysArrow keys, Tab, F1-F12, Home/End, Page Up/Down
Control CharactersCtrl-C, Ctrl-D, Ctrl-Z, Ctrl-L, telnet escape
Dangerous Command GateBlocks rm -rf, DROP TABLE, curl|sh - requires confirmation
OSC 133 Shell IntegrationAuto-detects command boundaries and exit codes
Smart TruncationFour strategies to prevent context overflow
Secret InputSend passwords without logging
Dynamic ResizeResize terminal on the fly with SIGWINCH
Idle CleanupAuto-closes idle sessions
Cross-PlatformLinux, macOS, and Windows support

Tools Reference

terminal-mcp exposes 9 MCP tools. Full details in docs/tools.md.

ToolPurpose
session_createSpawn a persistent terminal session
session_sendSend text, keys, or control characters
session_readRead output (stream, snapshot, auto, diff modes)
session_interactSend + read in one call
session_wait_forWait for regex pattern in output
session_execOne-shot command execution
session_closeClose a session gracefully
session_resizeResize terminal dimensions
session_listList active sessions

Architecture

flowchart LR
    Client[AI Client] -->|MCP JSON-RPC| Server[terminal-mcp]
    Server --> SM[Session Manager]
    SM --> S1[PTY 1: bash]
    SM --> S2[PTY 2: python3]
    SM --> S3[PTY 3: ssh user@host]
    S1 & S2 & S3 -.->|PTY output| Reader[Reader Thread]
    Reader -.->|buffer| Server

Each session is backed by a real PTY via pexpect.spawn (or PopenSpawn on Windows). For full architecture details, see docs/architecture.md.


Configuration

All settings configurable via TERMINAL_MCP_* environment variables. Full reference in docs/configuration.md.

SettingEnv VarDefault
Max sessionsTERMINAL_MCP_MAX_SESSIONS10
Idle timeoutTERMINAL_MCP_IDLE_TIMEOUT1800 (30 min)
Safety gateTERMINAL_MCP_SAFETY_GATEon
Buffer capTERMINAL_MCP_MAX_BUFFER_BYTES1000000 (1MB)
TruncationTERMINAL_MCP_TRUNCATION_MODEtail

Example with custom settings:

{
  "mcpServers": {
    "terminal": {
      "command": "uvx",
      "args": ["terminal-mcp"],
      "env": {
        "TERMINAL_MCP_MAX_SESSIONS": "20",
        "TERMINAL_MCP_IDLE_TIMEOUT": "3600",
        "TERMINAL_MCP_TRUNCATION_MODE": "head_tail"
      }
    }
  }
}

Documentation

DocumentDescription
Tools ReferenceComplete API for all 9 MCP tools
ArchitectureHow terminal-mcp works under the hood
ConfigurationAll settings and environment variables
Safety & SecurityDangerous command detection and safety gate
Use Cases & ExamplesReal-world recipes and patterns
ChangelogVersion history and release notes
ContributingHow to contribute

Supported Clients

ClientStatusInstall
Claude Code (CLI)Supported~/.claude.json or .mcp.json
Claude DesktopSupportedOne-click install
VS Code (Copilot Chat)SupportedOne-click install or .vscode/mcp.json
CursorSupportedOne-click install or Settings
WindsurfSupported~/.codeium/windsurf/mcp_config.json

Running Tests

pip install -e ".[dev]"
pytest tests/ -v

Contributing

Contributions welcome! See docs/contributing.md for guidelines.

License

MIT

Related MCP servers

Let AI agents buy dofollow backlinks: search 50,000+ vetted sites, order, publish. 7 hosted tools.

0
View repository →

Robinhood Chain intelligence: trend scores, launch radar, KOL leaderboard, pre-trade risk checks.

View repository →
AGAgent Ready logo

Scan any URL for AI agent readability — Vercel Spec, llmstxt.org, and agent-protocol manifests.

1
TypeScript
MIT
View repository →

Catch AI-fabricated citations and audit bibliographies with retraction checks, open-access lookup, and 10,000+ citation styles.

8
TypeScript
MIT
View repository →

Input/output safety for AI agents: known-pattern injection and obfuscation scan, URL/IP, secrets.

0
JavaScript
MIT
View repository →

All six agent guard suites in one local MCP server, with optional cloud intelligence.

0
JavaScript
MIT
View repository →