mcp-server-terminal
MCP server enabling AI agents to interact with terminal applications through structured Terminal State Tree representation. Works with any AI assistant that supports the Model Context Protocol.
README
Terminal MCP Server
MCP server enabling AI agents to interact with terminal applications through structured Terminal State Tree representation. Works with any AI assistant that supports the Model Context Protocol.
Installation
Via npm (Recommended)
npx mcp-server-terminal
Via GitHub Releases
Download pre-built binaries from Releases.
Build from Source
git clone https://github.com/aybelatchane/mcp-server-terminal.git
cd mcp-server-terminal
cargo build --release
# Binary: ./target/release/terminal-mcp
Configuration
Claude
Claude Code (CLI)
claude mcp add terminal -- npx mcp-server-terminal
Claude Desktop
Add to ~/.claude.json (macOS/Linux) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
OpenAI Codex
Codex CLI
codex mcp add terminal -- npx mcp-server-terminal
Codex Configuration File
Add to ~/.codex/config.toml:
[mcp_servers.terminal]
command = "npx"
args = ["mcp-server-terminal"]
Google Gemini
Gemini CLI
gemini mcp add terminal npx mcp-server-terminal
VS Code / GitHub Copilot
VS Code 1.101+ supports MCP. Add to your VS Code settings (settings.json):
{
"mcp.servers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
Cursor
Add to ~/.cursor/mcp.json or .cursor/mcp.json in your project:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
Zed
Add to your Zed settings (Preferences → Settings):
{
"context_servers": {
"terminal": {
"command": {
"path": "npx",
"args": ["mcp-server-terminal"]
}
}
}
}
Cline (VS Code Extension)
Click MCP Servers icon → Configure → Advanced MCP Settings, then add:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
AWS Bedrock
Add to your Bedrock agent MCP configuration:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"]
}
}
}
Other MCP Clients
For any MCP-compatible client, configure the server with:
- Command:
npx - Args:
["mcp-server-terminal"]
Or if using the binary directly:
- Command:
terminal-mcp
Usage
Ask your AI agent:
- "Create a terminal session running htop"
- "Take a snapshot of the terminal"
- "Press the down arrow key"
- "Type 'ls -la' and press Enter"
MCP Tools
| Tool | Description |
|---|---|
terminal_session_create |
Start a terminal session |
terminal_session_list |
List active sessions |
terminal_session_close |
Close a session |
terminal_session_resize |
Resize terminal dimensions |
terminal_snapshot |
Capture terminal state with UI elements |
terminal_type |
Type text into terminal |
terminal_press_key |
Press keys (arrows, F-keys, Ctrl+X) |
terminal_click |
Click on detected UI element |
terminal_wait_for |
Wait for text, element, or idle state |
terminal_read_output |
Read raw terminal output |
Visual Mode
By default, sessions spawn a visible terminal window (xterm). For headless operation:
npx mcp-server-terminal --headless
Or in your MCP config:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal", "--headless"]
}
}
}
X11 Setup (Linux/WSL)
Visual mode requires X11. Add the DISPLAY environment variable to your MCP config:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"],
"env": {
"DISPLAY": ":0"
}
}
}
}
Logging
Set the RUST_LOG environment variable:
{
"mcpServers": {
"terminal": {
"command": "npx",
"args": ["mcp-server-terminal"],
"env": {
"RUST_LOG": "info"
}
}
}
}
Log levels: error, warn, info, debug, trace
Logs go to stderr (stdout is reserved for MCP protocol).
Platform Support
| Platform | Architecture | Status |
|---|---|---|
| Linux | x64, arm64 | Supported |
| macOS | x64, arm64 | Supported |
| Windows | WSL | Supported |
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。