From Localhost to Production — 6 Steps

What You'll Build

Step 2: HTTP Deployment with FastMCP v3

Switch from stdio to HTTP with the ASGI app pattern — gives you control over workers, middleware, and process management:

# app.py
from fastmcp import FastMCP

mcp = FastMCP("ProductionServer")

@mcp.tool()
def search_docs(query: str) -> str:
    """Search internal documentation."""
    return f"Found results for: {query}"

# Create ASGI app for uvicorn/gunicorn
app = mcp.http_app()

uvicorn app:app --host 127.0.0.1 --port 8000 --workers 4

Or with gunicorn for process management:

gunicorn app:app -w 4 -k uvicorn.workers.UvicornWorker --bind 127.0.0.1:8000

Your MCP endpoint is now at http://127.0.0.1:8000/mcp.

Step 4: Authentication Layer

FastMCP v3 supports Bearer tokens, JWT, and OAuth. For team use, start with Bearer token auth:

from fastmcp.server.auth import BearerAuthProvider
import os

auth = BearerAuthProvider(
    token=os.environ["MCP_API_TOKEN"],
    jwt_secret=os.environ.get("JWT_SECRET"),  # Survives restarts
)

mcp = FastMCP("SecureServer", auth=auth)

Enterprise OAuth (Google/GitHub/Azure SSO)

from fastmcp.server.auth import OAuthProvider

auth = OAuthProvider(
    client_id=os.environ["OAUTH_CLIENT_ID"],
    client_secret=os.environ["OAUTH_CLIENT_SECRET"],
    authorize_url="https://accounts.google.com/o/oauth2/auth",
    token_url="https://oauth2.googleapis.com/token",
    jwt_secret=os.environ["JWT_SECRET"],
    token_store="redis://localhost:6379/0",
)

mcp = FastMCP("EnterpriseServer", auth=auth)

Step 6: Kubernetes Deployment

Use K8s when you need auto-scaling, rolling deployments, or multi-region. For most teams, Docker Compose is enough.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: fastmcp-server
spec:
  replicas: 3
  selector:
    matchLabels:
      app: fastmcp-server
  template:
    metadata:
      labels:
        app: fastmcp-server
    spec:
      containers:
        - name: fastmcp
          image: your-registry/fastmcp-server:latest
          ports:
            - containerPort: 8000
          env:
            - name: MCP_API_TOKEN
              valueFrom:
                secretKeyRef:
                  name: mcp-secrets
                  key: api-token
            - name: JWT_SECRET
              valueFrom:
                secretKeyRef:
                  name: mcp-secrets
                  key: jwt-secret
          readinessProbe:
            httpGet:
              path: /health
              port: 8000
            initialDelaySeconds: 5
            periodSeconds: 10
          resources:
            requests:
              memory: "256Mi"
              cpu: "250m"
            limits:
              memory: "512Mi"
              cpu: "500m"

Pair with an Ingress controller (nginx-ingress, Traefik) for TLS termination and routing.

FAQ