Fly.io MCP Server

About

## Fly.io MCP Server: Global Application Platform The **Fly.io MCP Server** integrates Fly.io into Google Antigravity, enabling global application deployment with edge computing, distributed databases, and GPU support directly from your development environment. ### Why Fly.io MCP? - **Global Edge**: Deploy to 30+ regions worldwide - **Machines API**: Fine-grained compute control - **Persistent Storage**: Volumes for stateful apps - **GPU Support**: ML workloads on GPUs - **Private Networking**: Secure app connectivity ### Key Features #### 1. App Deployment ```python # Deploy app deployment = await mcp.deploy( app="my-app", image="my-app:latest", regions=["iad", "lhr", "sin"], config={ "env": {"NODE_ENV": "production"}, "services": [{ "internal_port": 8080, "protocol": "tcp", "ports": [{"port": 443, "handlers": ["tls", "http"]}] }] } ) # Get app status app = await mcp.get_app(name="my-app") print(f"Status: {app['status']}") print(f"Hostname: {app['hostname']}") ``` #### 2. Machine Management ```python # Create machine machine = await mcp.create_machine( app="my-app", region="iad", config={ "image": "my-app:latest", "size": "shared-cpu-1x", "env": {"ROLE": "worker"} } ) # List machines machines = await mcp.list_machines(app="my-app") for m in machines: print(f"Machine: {m['id']} - {m['region']} - {m['state']}") # Stop machine await mcp.stop_machine(app="my-app", machine_id=machine["id"]) ``` #### 3. Volume Operations ```python # Create volume volume = await mcp.create_volume( app="my-app", name="data", region="iad", size_gb=10 ) # Attach to machine await mcp.update_machine( app="my-app", machine_id="machine_123", config={ "mounts": [{ "volume": volume["id"], "path": "/data" }] } ) ``` #### 4. Secrets Management ```python # Set secrets await mcp.set_secrets( app="my-app", secrets={ "DATABASE_URL": "postgres://...", "API_KEY": "secret-key" } ) # List secrets secrets = await mcp.list_secrets(app="my-app") ``` ### Configuration ```json { "mcpServers": { "fly-io": { "command": "npx", "args": ["-y", "@anthropic/mcp-fly"], "env": { "FLY_API_TOKEN": "your-api-token" } } } } ``` ### Use Cases **Global Apps**: Deploy apps close to users worldwide. **Stateful Services**: Run databases and caches at edge. **ML Inference**: Deploy models on edge GPUs. **Real-Time**: Low-latency real-time applications. The Fly.io MCP enables global deployment within your development environment.

Installation

{
  "mcpServers": {
    "fly-io": {
      "mcpServers": {
        "fly-io": {
          "env": {
            "FLY_API_TOKEN": "your-api-token"
          },
          "args": [
            "-y",
            "fly-mcp-server"
          ],
          "command": "npx"
        }
      }
    }
  }
}

How to Use

1. 1Get Fly.io API token
2. 2Configure environment
3. 3Deploy globally