Tunio MCP lets your AI assistant run your radio stations for you. It works with everything the Tunio public API offers: stations, content and music schedules, playlists, jingles and jingle bundles, announcements, TTS, podcasts, Tunio Stories playlists and air history. You ask in plain words, for example "what is playing on my stations right now?", "add the Morning playlist to Monday's schedule" or "make a playlist of English fairy tales for kids under 6 and play three of them at random every evening at 8", and the assistant calls the right tools.
The server address is the same for every client:
https://mcp.tunio.ai/mcpYou sign in with your Tunio account. No API key and no extra software are needed.
What the assistant can do
On the first connection Tunio shows a consent screen, where you allow two kinds of access:
- View — your stations, schedules, playlists, story playlists, jingles and air history.
- Manage — change schedules, playlists, story playlists, jingles, announcements and playback.
The assistant always acts as you and sees only your account. Tools that change or delete data are marked as such, so the client asks you before running them. The assistant can also generate audio: new jingles, TTS and voiced announcements. Generation spends your account's credits.
The consent screen marks AI clients as third-party apps. Check the address after "After you allow access you will be sent to" (for Claude it is claude.ai) and allow access only to apps you trust.
Connect in ChatGPT
Tunio is in the ChatGPT app directory, so you don't need a custom connector or developer mode.
Open Tunio in ChatGPT
Follow the link to Tunio in ChatGPT or search for "Tunio" in the ChatGPT app directory.
Connect and sign in with Tunio
Click the connect button. ChatGPT opens the Tunio sign-in page: sign in, check the requested access and click Allow.
Ask for something
Type in the chat, for example, "show my Tunio stations". Before anything that changes your data, ChatGPT asks you to confirm.
Connect in Claude (web and desktop)
Add a custom connector
Open Settings → Connectors and click Add custom connector. Name it
Tunioand pastehttps://mcp.tunio.ai/mcpas the URL.Sign in with Tunio
Click Connect. Claude opens the Tunio sign-in page: sign in, check the requested access and click Allow.
Choose tool permissions
Back in Claude, the Tunio connector lists its tools. By default Claude asks before each call; you can allow the read-only tools once and keep approval for the rest.
Connect in Claude Code
Add the server
Run in a terminal:
BASHclaude mcp add --transport http tunio https://mcp.tunio.ai/mcpSign in
Start Claude Code, run
/mcp, choose tunio and Authenticate. Sign in with Tunio in the browser window and allow access.
Connect in Cursor
Add the server to mcp.json
Open Cursor Settings → MCP (or edit
~/.cursor/mcp.json) and add:JSON{ "mcpServers": { "tunio": { "url": "https://mcp.tunio.ai/mcp" } } }Sign in
Cursor shows that the server needs authentication. Click Login, sign in with Tunio and allow access.
Connect in other clients
Any MCP client that supports remote servers (Streamable HTTP) with OAuth connects the same way: use https://mcp.tunio.ai/mcp as the server URL and choose OAuth when the client asks how to authenticate.
Connect with an API key
If your client cannot sign in with OAuth, pass an API key instead. Create one at cp.tunio.ai/api-keys and send it as a header:
Authorization: Bearer YOUR_API_KEYFor a client that only runs local (stdio) servers, connect through mcp-remote:
npx -y mcp-remote https://mcp.tunio.ai/mcp --header "Authorization: Bearer YOUR_API_KEY"An API key gives full access to your account. Never put it in screenshots, repositories or chats. If a key leaks, delete it and create a new one.
Check the connection
Ask the assistant something simple, for example "list my Tunio stations" or "what is playing on my station now?". If it answers with your data, Tunio MCP is connected.
If something does not work:
- The tools do not appear. Remove the connector and add it again, then restart the client.
- "This action changes data and needs the api:write permission." The connection has view access only. Disconnect Tunio in the client and connect again.
To stop the assistant from using your account, remove the Tunio connector in your AI client.