python_local MCP Server

What is python_local MCP Server?

The python_local MCP Server provides an interactive Python REPL (Read-Eval-Print Loop) environment as an MCP server. It is designed for use with AI assistants like Claude, allowing them to execute Python code and retrieve session history through the Model Context Protocol.

How to use python_local MCP Server?

Install the server via uvx or configure it locally using uv in your Claude Desktop configuration file (e.g., claude_desktop_config.json). Use the python_repl tool with the required code and session_id arguments to execute Python code in a persistent session.

Key features of python_local MCP Server

• Provides an interactive Python REPL via MCP
• Exposes repl:// resource for session history
• Executes Python code with the python_repl tool
• Maintains separate state per session
• Supports both expressions and statements
• Captures stdout and stderr output

Use cases of python_local MCP Server

• Execute Python code interactively from an AI assistant
• Maintain separate state for multiple coding sessions
• View the history of executed code and output via resources
• Debug or test small Python snippets within a conversation

FAQ from python_local MCP Server

What does python_local MCP Server do?

It provides an interactive Python REPL environment accessible through the MCP protocol, enabling AI assistants to run Python code and retrieve session history.

What tools does it provide?

The server provides one tool, python_repl, which takes code (Python code to execute) and session_id (to identify the session) as required arguments.

How is session history accessed?

Session history is available via the custom repl:// URI scheme, which returns the history as text/plain, showing input code and corresponding output for each execution.

How are sessions managed?

Each session is identified by a session_id and maintains separate state, allowing multiple independent REPL sessions to run concurrently.

How can I debug the server?

You can launch the MCP Inspector via npx @modelcontextprotocol/inspector uv --directory /path/to/python_local run python-local and access the debugging URL in your browser.