概览
What is MCP Interactive Terminal Server?
The MCP Interactive Terminal Server allows an AI agent to interact with command-line driven prompts, such as gdb, a UART console, or any REPL-based tool. Built with FastMCP, it provides tools for managing interactive terminal sessions and supports stdio, sse, and streamable-http transports.
How to use MCP Interactive Terminal Server?
Clone the repository, ensure Python 3.12+ and optionally UV Astral are installed, then run the server via the command line (e.g., uv run python -m interactive.server) or configure it in VS Code’s .vscode/mcp.json and start it through the MCP interface.
Key features of MCP Interactive Terminal Server
- Start and manage interactive terminal sessions.
- Wait for specific output or prompts with timeouts.
- Send commands to active sessions.
- Terminate sessions on demand.
- Retrieve a list of all active sessions.
Use cases of MCP Interactive Terminal Server
- Automating debugging sessions with gdb.
- Interacting with UART console output.
- Managing any REPL-based tool programmatically.
- Running interactive Python sessions from an agent.
FAQ from MCP Interactive Terminal Server
What are the runtime prerequisites for MCP Interactive Terminal Server?
Python 3.12 or later is required. The UV Astral package manager is optional but recommended for handling dependencies.
What transport options does the server support?
The server supports stdio (default), sse, and streamable-http transports, selectable via command-line arguments.
Can I log the output of an interactive session?
Yes, the start_session tool accepts an optional log_file parameter to write all session output to a specified file.
How do I start a new interactive session?
Use the start_session tool with a command string and optional arguments list, e.g., start_session("python", ["-i"]) starts a Python REPL.
How do I retrieve active sessions?
Use the get_active_sessions tool to obtain a dictionary mapping session IDs to their current status.
