Python A decorator, 3 imports,
A decorator, 3 imports,
and a runtime you didn't ask for.
zeromcp vs FastMCP — side by side.
Dependencies
0 ZeroMCP
vs. 37 Official SDK
Throughput
3.08K req/s ZeroMCP
vs. 623 req/s Official SDK
Memory
27 MB ZeroMCP
vs. 87 MB Official SDK
This is a hello world
A decorator, a server class, and a run call. The official SDK wraps everything in ceremony before your tool does anything.
FastMCP 8 lines
from mcp.server.fastmcp import FastMCP
mcp = FastMCP("test")
@mcp.tool()
def hello(name: str) -> str:
return f"Hello, {name}!"
mcp.run()This is the whole server
No decorator. No server class. No run call.
Drop it in a folder and run zeromcp serve.
ZeroMCP 7 lines
# tools/hello.py
tool = {
"description": "Say hello",
"input": {"name": "string"},
}
async def execute(args, ctx):
return f"Hello, {args['name']}!"HTTP Performance — Head to Head
Mixed workload across all 7 MCP method types. 5-minute sustained load in Docker. Starlette for ZeroMCP, stdio proxy for the official SDK.
req/s CPU Memory Ratio
ZeroMCP (Starlette) 3.08K 0.28% 27 MB 4.9x
Official SDK 623 0.04% 87 MB
The tradeoff
Choose FastMCP
If you want the decorator-based API and tight integration with pydantic validation. It's the most Pythonic MCP experience.
- Type hints — Pydantic validation integrated via FastMCP
- Spec parity — tracks every spec change immediately
- Enterprise support — maintained by the MCP specification team at Anthropic
Choose ZeroMCP
If you want zero dependencies and production HTTP performance. Drop a .py file, serve on Starlette or FastAPI.
- 0 dependencies
- File-based tools — drop a .py file, it's live
- Starlette + FastAPI + Flask
- Built-in sandbox with enforced permissions
- 3.08K req/s on Starlette