MCP Code Execute (Python)

Executes Python code in a sandboxed environment using Pyodide (Python compiled to WebAssembly).

• Tech: TypeScript, Deno, Pyodide (Python 3.12 in WASM)
• Port: 8000
• Tool name: run_python_code

Tool: run_python_code

Runs Python code in an isolated sandbox. Supports async Python, dependency management, and file creation.

Input Schema

Field	Type	Required	Description
code	string	Yes	Python code to execute

Example

{
  "name": "run_python_code",
  "arguments": {
    "code": "import math\nresult = math.sqrt(144)\nprint(f'Square root: {result}')\nresult"
  }
}

Response

{
  "stdout": "Square root: 12.0\n",
  "stderr": "",
  "result": 12.0,
  "files": []
}

Features

Dependency Management

Dependencies can be declared using PEP 723 script metadata in the code:

# /// script
# dependencies = ["numpy", "pandas"]
# ///
import numpy as np
import pandas as pd

data = np.array([1, 2, 3, 4, 5])
print(f"Mean: {np.mean(data)}")

Pyodide installs the packages via micropip before execution.

File Creation

Code can create files in /tmp/mcp_run_python. After execution, all created files are automatically uploaded to Azure Blob Storage and their URLs are returned.

import matplotlib.pyplot as plt

plt.plot([1, 2, 3], [4, 5, 6])
plt.savefig('/tmp/mcp_run_python/chart.png')

Security

• Code runs in WebAssembly (Pyodide) -- no access to the host filesystem or network
• Execution is sandboxed by design
• File output is restricted to the /tmp/mcp_run_python directory

Transports

Supports three MCP transport modes: stdio, SSE, and Streamable HTTP.