Goose FM – MCP Server for Real‑Time Radio Streaming

Goose FM demonstrates how an MCP server can turn a simple hardware setup into a fully‑integrated AI tool. By connecting a cheap RTL‑SDR dongle and an antenna to a computer, the server allows an AI assistant (e.g., Claude) to tune into any FM broadcast and stream it directly through the user’s speakers. The problem it solves is two‑fold: first, it abstracts the low‑level radio hardware and software stack into a single, discoverable MCP service; second, it gives conversational AI the ability to control and stream live radio content without leaving the chat interface.

The server exposes a minimal yet powerful set of MCP capabilities. It offers resources that describe the available radio stations, a tool that accepts station identifiers or frequencies and initiates tuning, and a prompt that can be used to guide the AI in selecting stations or adjusting volume. Under the hood, Goose FM orchestrates to capture the radio signal and pipes it through for audio output, all while keeping the MCP protocol transparent to the client. This design means developers can plug Goose FM into any AI workflow that supports MCP, whether it’s a desktop assistant, a chatbot on a website, or an automated voice system.

Key features include:

• Live FM tuning: The AI can specify frequencies or station names, and the server handles all signal processing.
• Audio streaming: Tuned audio is streamed to the client’s speakers in real time, enabling hands‑free listening.
• Simple integration: Adding Goose FM to a Claude desktop or any MCP‑compliant client requires only a short JSON snippet, with no need to manage drivers or low‑level libraries manually.
• Extensibility: Because the server follows MCP standards, developers can extend it with additional controls (e.g., volume, recording) or integrate it into larger media pipelines.

Real‑world use cases are plentiful. In a smart home, an AI assistant could play news or music from local stations when asked to “turn on the radio.” In a museum exhibit, visitors could request specific historical broadcasts via voice. For developers, Goose FM serves as a low‑cost, hardware‑based demo of how MCP can bridge physical devices with conversational AI, showcasing the protocol’s flexibility and the ease of turning niche hardware into a consumable AI service.