fastmcp

What this skill does

# FastMCP - Build MCP Servers in Python

FastMCP is a Python framework for building Model Context Protocol (MCP) servers that expose tools, resources, and prompts to Large Language Models like Claude.

## Quick Start

### Installation

```bash
pip install fastmcp
# or: uv pip install fastmcp
```

### Minimal Server

```python
from fastmcp import FastMCP

# MUST be at module level for FastMCP Cloud
mcp = FastMCP("My Server")

@mcp.tool()
async def hello(name: str) -> str:
    """Say hello to someone."""
    return f"Hello, {name}!"

if __name__ == "__main__":
    mcp.run()
```

**Run:**
```bash
python server.py              # Local development
fastmcp dev server.py         # With FastMCP CLI
python server.py --transport http --port 8000  # HTTP mode
```

**Copy-Paste Template**: See `templates/basic-server.py`

## Core Concepts

### Tools

Functions that LLMs can call:

```python
@mcp.tool()
def calculate(operation: str, a: float, b: float) -> float:
    """Perform mathematical operations."""
    operations = {
        "add": lambda x, y: x + y,
        "subtract": lambda x, y: x - y,
        "multiply": lambda x, y: x * y,
        "divide": lambda x, y: x / y if y != 0 else None
    }
    return operations.get(operation, lambda x, y: None)(a, b)
```

**Best Practices:**
- Clear, descriptive function names
- Comprehensive docstrings (LLMs read these!)
- Strong type hints (Pydantic validates automatically)
- Return structured data (dicts/lists)
- Handle errors gracefully

### Resources

Expose static or dynamic data:

```python
@mcp.resource("data://config")
def get_config() -> dict:
    """Provide application configuration."""
    return {"version": "1.0.0", "features": ["auth", "api"]}

# Dynamic resource with parameters
@mcp.resource("user://{user_id}/profile")
async def get_user_profile(user_id: str) -> dict:
    """Get user profile by ID."""
    return {"id": user_id, "name": f"User {user_id}"}
```

**URI Schemes**: `data://`, `file://`, `resource://`, `info://`, `api://`, or custom

### Prompts

Pre-configured prompts for LLMs:

```python
@mcp.prompt("analyze")
def analyze_prompt(topic: str) -> str:
    """Generate analysis prompt."""
    return f"""Analyze {topic} considering:
    1. Current state
    2. Challenges
    3. Opportunities
    4. Recommendations"""
```

### Context Features

**Progress Tracking:**
```python
from fastmcp import Context

@mcp.tool()
async def batch_process(items: list, context: Context) -> dict:
    """Process items with progress updates."""
    for i, item in enumerate(items):
        await context.report_progress(i + 1, len(items), f"Processing {item}")
        await process_item(item)
    return {"processed": len(items)}
```

**User Input:**
```python
@mcp.tool()
async def confirm_action(action: str, context: Context) -> dict:
    """Perform action with user confirmation."""
    confirmed = await context.request_elicitation(
        prompt=f"Confirm {action}? (yes/no)",
        response_type=str
    )
    return {"confirmed": confirmed.lower() == "yes"}
```

## Storage Backends

Choose storage based on deployment:

```python
from key_value.stores import DiskStore, RedisStore
from key_value.encryption import FernetEncryptionWrapper
from cryptography.fernet import Fernet

# Memory (default) - Development only
mcp = FastMCP("Dev Server")

# Disk - Single instance
mcp = FastMCP(
    "Production Server",
    storage=FernetEncryptionWrapper(
        key_value=DiskStore(path="/var/lib/mcp/storage"),
        fernet=Fernet(os.getenv("STORAGE_ENCRYPTION_KEY"))
    )
)

# Redis - Multi-instance
mcp = FastMCP(
    "Production Server",
    storage=FernetEncryptionWrapper(
        key_value=RedisStore(
            host=os.getenv("REDIS_HOST"),
            password=os.getenv("REDIS_PASSWORD")
        ),
        fernet=Fernet(os.getenv("STORAGE_ENCRYPTION_KEY"))
    )
)
```

## Server Lifespans

Initialize resources on server startup:

```python
from contextlib import asynccontextmanager

@asynccontextmanager
async def app_lifespan(server: FastMCP):
    """Runs ONCE when server starts (v2.13.0+)."""
    db = await Database.connect()
    print("Server starting")
    
    try:
        yield {"db": db}
    finally:
        await db.disconnect()
        print("Server stopping")

mcp = FastMCP("My Server", lifespan=app_lifespan)
```

**Critical**: v2.13.0+ lifespans run per-server (not per-session). For per-session logic, use middleware.

## Middleware System

8 built-in middleware types:

```python
from fastmcp.middleware import (
    LoggingMiddleware,
    TimingMiddleware,
    RateLimitingMiddleware,
    ResponseCachingMiddleware
)

# Order matters!
mcp.add_middleware(LoggingMiddleware())
mcp.add_middleware(TimingMiddleware())
mcp.add_middleware(RateLimitingMiddleware(max_requests=100, window_seconds=60))
mcp.add_middleware(ResponseCachingMiddleware(ttl_seconds=3600))
```

**Custom Middleware:**
```python
from fastmcp.middleware import BaseMiddleware

class CustomMiddleware(BaseMiddleware):
    async def on_call_tool(self, tool_name, arguments, context):
        print(f"Before: {tool_name}")
        result = await self.next(tool_name, arguments, context)  # MUST call next()
        print(f"After: {tool_name}")
        return result
```

## Server Composition

**Import Server** (static, one-time copy):
```python
main_server.import_server(vendor_server)  # Static bundle
```

**Mount Server** (dynamic, runtime delegation):
```python
main_server.mount(api_server, prefix="api")  # Changes appear immediately
```

## Cloud Deployment

**FastMCP Cloud Requirements:**
1. Server MUST be at module level
2. Use disk/Redis storage (not memory)
3. No import-time execution

```python
# ✅ Cloud-ready pattern
mcp = FastMCP("My Server")  # Module level

@mcp.tool()
async def my_tool(): pass

if __name__ == "__main__":
    mcp.run()
```

**Deploy:**
```bash
fastmcp deploy server.py
```

## Top 5 Critical Errors

### 1. Missing Server Object

**Error:** `RuntimeError: No server object found at module level`

**Fix:**
```python
# ❌ WRONG
def create_server():
    return FastMCP("server")

# ✅ CORRECT
mcp = FastMCP("server")  # At module level
```

### 2. Async/Await Confusion

**Error:** `RuntimeError: no running event loop`

**Fix:**
```python
# ❌ WRONG: Sync function calling async
@mcp.tool()
def bad_tool():
    result = await async_function()  # Error!

# ✅ CORRECT: Async tool
@mcp.tool()
async def good_tool():
    result = await async_function()
    return result
```

### 3. Context Not Injected

**Error:** `TypeError: missing 1 required positional argument: 'context'`

**Fix:**
```python
from fastmcp import Context

# ❌ WRONG: No type hint
@mcp.tool()
async def bad_tool(context):  # Missing type!
    await context.report_progress(...)

# ✅ CORRECT: Proper type hint
@mcp.tool()
async def good_tool(context: Context):
    await context.report_progress(0, 100, "Starting")
```

### 4. Storage Backend Not Configured

**Error:** `RuntimeError: OAuth tokens lost on restart`

**Fix:** Use disk or Redis storage in production (see Storage Backends section above)

### 5. Circular Import Errors

**Error:** `ImportError: cannot import name 'X' from partially initialized module`

**Fix:**
```python
# ❌ WRONG: Factory function creating circular dependency
# shared/__init__.py
def get_client():
    from .api_client import APIClient  # Circular!
    return APIClient()

# ✅ CORRECT: Direct imports
# shared/__init__.py
from .api_client import APIClient
from .cache import CacheManager

# shared/monitoring.py
from .api_client import APIClient
client = APIClient()
```

**See all 25 errors**: `references/error-catalog.md`

## Client Configuration

### Claude Desktop

```json
{
  "mcpServers": {
    "my-server": {
      "command": "python",
      "args": ["/path/to/server.py"]
    }
  }
}
```

### Claude Code CLI

```json
{
  "mcpServers": {
    "my-server": {
      "command": "python",
      "args": ["server.py"]
    }
  }
}
```

## CLI Commands

```bash
fastmcp dev serv