respan-instrumentation-mcp

Respan instrumentation plugin for Model Context Protocol Python applications.

The package enables upstream OpenInference MCP transport context propagation and adds Respan spans for common mcp.ClientSession operations such as tool listing, tool calls, resource reads, and prompt fetches.

Configuration

1. Install

pip install respan-ai respan-instrumentation-mcp "mcp>=1.27,<2"

2. Set Environment Variables

Quickstart

import asyncio
import sys

from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
from respan import Respan, workflow
from respan_instrumentation_mcp import MCPInstrumentor

respan = Respan(instrumentations=[MCPInstrumentor()])


@workflow(name="mcp_tool_call_workflow")
async def run_mcp_client() -> None:
    server_params = StdioServerParameters(
        command=sys.executable,
        args=["server.py"],
    )

    async with stdio_client(server_params) as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()
            tools = await session.list_tools()
            print([tool.name for tool in tools.tools])
            result = await session.call_tool(
                "summarize_city",
                arguments={"city": "Paris"},
            )
            print(result)


asyncio.run(run_mcp_client())
respan.flush()
respan.shutdown()

This release supports MCP Python SDK 1.27 or newer on the 1.x line. MCP 2.x removed the mcp.server.fastmcp API used by the current examples and is not yet supported.

Further Reading

See the Respan example projects for runnable MCP scripts.