chatgpt-api/docs/publishing/guides/py-fastmcp.mdx

---
title: Python FastMCP
description: This guide will show you how to publish an MCP server to Agentic using the Python fastmcp package.
---

[Python's FastMCP](https://gofastmcp.com) is a popular [open source](https://github.com/jlowin/fastmcp) MCP server framework for Python.

## 1. Install the `fastmcp` package

<Info>
  **Prerequisite**: Please install [Python](https://www.python.org) before
  proceeding.
</Info>

<CodeGroup>

```bash uv
uv add fastmcp
```

```bash pip
pip install fastmcp
```

</CodeGroup>

## 2. Create a FastMCP server

```py server.py
from fastmcp import FastMCP

mcp = FastMCP("Demo 🚀")

@mcp.tool
def add(a: int, b: int) -> int:
    """Add two numbers"""
    return a + b

if __name__ == "__main__":
    mcp.run(transport="http", host="127.0.0.1", port=8000, path="/mcp")
```

<Note>
  Make sure to run your server with the `transport="http"` option to use the
  Streamable HTTP transport.
</Note>

## 3. Deploy your MCP server remotely

Deploy your server publicly or use a tool like [ngrok](https://ngrok.com) to expose it to the internet.

<Warning>
  Tools like `ngrok` expose your unauthenticated server to the internet. Only
  run this command in a safe environment if you understand the risks.
</Warning>

We recommend deploying your server to a cloud provider like [render](https://render.com/docs/deploy-fastapi), [platform.sh](https://docs.platform.sh/languages/python.html), [porter](https://docs.porter.run/guides/fastapi/deploy-fastapi), or [fly.io](https://fly.io/docs/python/frameworks/fastapi/). Or one of the big boys [AWS](https://aws.amazon.com), [GCP](https://cloud.google.com), or [Azure](https://azure.microsoft.com).

## 4. Deploy your origin MCP server to Agentic

Now that you have a publicly available MCP server, you can follow the [existing MCP server guide](/publishing/guides/existing-mcp-server) to deploy it to Agentic.