MCP (Model Context Protocol) allows communication between external tools and AI-Client (here Claude-AI). This project sets up an MCP-compatible terminal server to execute shell commands within a defined workspace.
uv
Run this command in PowerShell:
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
If you have Python installed:
pip install uv # or
pipx install uv
After installation, restart your terminal so uv
is recognized.
Run the following commands in PowerShell:
mkdir D:\path\to\mcp\servers\terminal_server
mkdir $env:USERPROFILE\mcp\workspace
D:\path\to\mcp\servers\terminal_server
: Stores all MCP servers.$env:USERPROFILE\mcp\workspace
: Dedicated workspace directory.cd D:\path\to\mcp\servers\terminal_server
uv init
This initializes a Python project inside terminal_server
.
uv venv
.venv\Scripts\activate
This creates and activates a virtual environment, keeping dependencies isolated.
uv add "mcp[cli]"
This installs the MCP package to enable communication with Claude.
uv run terminal_server.py
This starts the terminal server within the virtual environment.
Edit the Claude config file at:
C:\Users\<your-username>\AppData\Roaming\Claude\claude_desktop_config.json
Add the following configuration:
{
"mcpServers": {
"terminal": {
"command": "C:\\path\\to\\uv.exe",
"args": [
"--directory", "D:\\path\\to\\mcp\\servers\\terminal_server",
"run",
"terminal_server.py"
]
}
}
}
Once restarted, you should see a hammer icon, meaning your tool is ready to use! 🚀
{
"mcpServers": {
"terminal": {
"env": {},
"args": [
"--directory",
"D:\\path\\to\\mcp\\servers\\terminal_server",
"run",
"terminal_server.py"
],
"command": "uv"
}
}
}
Seamless access to top MCP servers powering the future of AI integration.