fastmcp/docs/getting-started/quickstart.mdx
2025-05-16 11:22:55 +09:00

129 lines
4 KiB
Text

---
title: Quickstart
icon: rocket-launch
---
Welcome! This guide will help you quickly set up FastMCP and run your first MCP server.
If you haven't already installed FastMCP, follow the [installation instructions](/getting-started/installation).
## Creating a FastMCP Server
A FastMCP server is a collection of tools, resources, and other MCP components. To create a server, start by instantiating the `FastMCP` class.
Create a new file called `my_server.py` and add the following code:
```python my_server.py
from fastmcp import FastMCP
mcp = FastMCP("My MCP Server")
```
That's it! You've created a FastMCP server, albeit a very boring one. Let's add a tool to make it more interesting.
## Adding a Tool
To add a tool that returns a simple greeting, write a function and decorate it with `@mcp.tool` to register it with the server:
```python my_server.py {5-7}
from fastmcp import FastMCP
mcp = FastMCP("My MCP Server")
@mcp.tool()
def greet(name: str) -> str:
return f"Hello, {name}!"
```
## Testing the Server
To test the server, create a FastMCP client and point it at the server object.
```python my_server.py {1-2, 10-17}
import asyncio
from fastmcp import FastMCP, Client
mcp = FastMCP("My MCP Server")
@mcp.tool()
def greet(name: str) -> str:
return f"Hello, {name}!"
client = Client(mcp)
async def call_tool(name: str):
async with client:
result = await client.call_tool("greet", {"name": name})
print(result)
asyncio.run(call_tool("Ford"))
```
There are a few things to note here:
- Clients are asynchronous, so we need to use `asyncio.run` to run the client.
- We must enter a client context (`async with client:`) before using the client. You can make multiple client calls within the same context.
## Running the server
In order to run the server with Python, we need to add a `run` statement to the `__main__` block of the server file.
```python my_server.py {9-10}
from fastmcp import FastMCP
mcp = FastMCP("My MCP Server")
@mcp.tool()
def greet(name: str) -> str:
return f"Hello, {name}!"
if __name__ == "__main__":
mcp.run()
```
This lets us run the server with `python my_server.py`, using the default `stdio` transport, which is the standard way to expose an MCP server to a client.
<Tip>
Why do we need the `if __name__ == "__main__":` block?
Within the FastMCP ecosystem, this line may be unnecessary. However, including it ensures that your FastMCP server runs for all users and clients in a consistent way and is therefore recommended as best practice.
</Tip>
### Interacting with the Python server
Now that the server can be executed with `python my_server.py`, we can interact with it like any other MCP server.
In a new file, create a client and point it at the server file:
```python my_client.py
import asyncio
from fastmcp import Client
client = Client("my_server.py")
async def call_tool(name: str):
async with client:
result = await client.call_tool("greet", {"name": name})
print(result)
asyncio.run(call_tool("Ford"))
```
### Using the FastMCP CLI
To have FastMCP run the server for us, we can use the `fastmcp run` command. This will start the server and keep it running until it is stopped. By default, it will use the `stdio` transport, which is a simple text-based protocol for interacting with the server.
```bash
fastmcp run my_server.py:mcp
```
Note that FastMCP *does not* require the `__main__` block in the server file, and will ignore it if it is present. Instead, it looks for the server object provided in the CLI command (here, `mcp`). If no server object is provided, `fastmcp run` will automatically search for servers called "mcp", "app", or "server" in the file.
<Tip>
We pointed our client at the server file, which is recognized as a Python MCP server and executed with `python my_server.py` by default. This executes the `__main__` block of the server file. There are other ways to run the server, which are described in the [server configuration](/servers/fastmcp#running-the-server) guide.
</Tip>