mcpmini

Usage

Installation

pip install mcpmini

How to use

An MCP tool is a docmented Python function — the docments become the tool’s schema, with no registration ceremony. Serve some functions over stdio or HTTP:

from mcpmini.core import MCPServer, MCPClient, serve_stdio, serve_mcp
import asyncio, socket

def fahrenheit(
    celsius:float, # Temperature to convert
)->float:
    "Convert Celsius to Fahrenheit"
    return celsius*9/5+32

srv = MCPServer('demo', [fahrenheit])

asyncio.run(serve_stdio(srv)) serves it as a host-launched stdio server; a tool on that transport can call await srv.elicit(message, schema), and an MCPClient.stdio(..., on_request=handler) supplies the structured answer while the original tool call remains active. asyncio.run(serve_mcp(srv, port=8000, token='S')) serves it over streamable HTTP behind a bearer token (non-loopback binds refuse to start tokenless unless you pass no_token=True). A file of docmented functions serves straight from the command line — mcpmini tools.py, or mcpmini tools.py --transport http --port 8000 — which is the shape MCP host configs launch. Then any MCP client can connect — for example Claude Code:

claude mcp add --transport http demo http://127.0.0.1:8000/mcp -H "Authorization: Bearer S"

And mcpmini is a client too: server tools come back as bound Python functions, with their signatures, docs, and defaults rebuilt from the wire schema.

def free_port():
    with socket.socket() as s:
        s.bind(('127.0.0.1', 0))
        return s.getsockname()[1]
port = free_port()
task = asyncio.create_task(serve_mcp(srv, port=port, token='S'))
await asyncio.sleep(0.2)
async with MCPClient.http(f'http://127.0.0.1:{port}/mcp', token='S') as c: res = await c.tools.fahrenheit(celsius=100)
task.cancel()
res

'212.0'