Announcement · July 2026
The first radio station
with an MCP server.
KAZM is opening its live data to every AI assistant on earth — now playing, fire restrictions, road conditions, emergency alerts, and more, queryable in real time. No other station has done this.
How a mountain radio station got ahead of the networks
When most people think about a radio station’s technology, they picture a mixing board, a transmitter tower, maybe a streaming player on the website. KAZM has all of that. But it also runs something unusual for a small-market station: a live data infrastructure that’s been quietly powering one of the most feature-rich independent radio websites in the country.
Fire restriction levels from the Coconino National Forest, refreshed every six hours. Road conditions pulled from ADOT’s live feeds. Sedona weather straight from National Weather Service coordinates at the transmitter site. Concert tour dates for every artist in rotation, deduplicated and verified. A 14-day rewind archive, block by block. A song history going back to October 2023 that can answer what was playing at any minute since then.
All of that already exists. The MCP server simply opens a door into it.
What MCP is, in plain terms
Model Context Protocol is Anthropic’s open standard for connecting AI assistants to live data and tools. Think of it as a USB port — but for knowledge. When an AI assistant needs real-world information to answer a question, it queries an MCP server. The server responds. The AI tells you, with the source intact.
“Right now, if you ask an AI ‘what are the fire restrictions in Sedona today?’ it either guesses, searches, or admits it doesn’t know. With KAZM’s MCP server, it knows.”
The protocol is model-agnostic and already supported by Claude and ChatGPT, with adoption moving fast across the AI ecosystem. A station that has an MCP server today is positioned the way a station with a strong website was positioned in 2002 — early enough that being there matters.
45 live tools, all real data
Every tool reads from a live source. Nothing is cached longer than the underlying data warrants. Nothing is fabricated.
with MCP servers
and growing
on the mountain
October of that year
Search indexes what you publish.
AI queries what you make available.
These are not the same thing, and the distinction will matter more with every passing year.
A page on a website gets discovered when someone searches for it. An MCP tool gets called when an AI needs it — which is increasingly whenever anyone asks a question. The shift from “I’ll Google it” to “I’ll ask Claude” is already in progress, and most media organizations haven’t thought about what that means for their distribution.
No other radio station — not iHeart, not NPR, not the big-market FM flagships with full engineering teams — has done this. That’s not an indictment of them. It’s an opportunity for KAZM. Being first in a distribution channel has historically been worth more than being second with a bigger budget.
And there’s a specifically local dimension here. When an AI helps someone plan a Sedona trip, or check road conditions before heading up Oak Creek Canyon, or ask about fire restrictions before lighting a campfire — KAZM is the answer. Real local data, properly attributed, in AI conversations instead of aggregator sites that scraped it from somewhere else.
Add KAZM to your AI assistant
Do this once and your AI can answer questions like “What’s playing on KAZM?” or “Are there fires near Sedona today?” with live, real answers. Pick your app below.
Claude Desktop
You need the Claude Desktop app on your Mac or Windows PC (not the website — the actual app). Get it free here if you don’t have it yet. Then click the button below — Claude opens, asks you to confirm, and you’re done. After that, just ask Claude anything about KAZM and it will have live answers.
Add to Claude DesktopChatGPT
Works in the ChatGPT app on your computer (Windows or Mac) — not the website. Takes about two minutes, and you only do it once. After that, ask ChatGPT anything about KAZM and it will know.
- Open the ChatGPT app on your computer Don’t have it yet? Download it free here →
- Open Settings (click your name or the gear icon ⚙), then find Connectors or Integrations The exact name depends on your version — look for something about connected apps or plugins.
- Click “Add” or “Add MCP Server” and paste the address you copied in Step 1 To paste: click in the box, then press Ctrl+V (Windows) or ⌘+V (Mac).
- Name it KAZM and click Save Done. KAZM’s live data is now in every ChatGPT conversation.
https://mcp.mellowmountainradio.com/mcp
Cursor, Windsurf, Continue & any other MCP client
If your AI tool supports MCP, add KAZM with one line of config. Copy the snippet below and paste it into your client’s MCP server settings file.
{
"mcpServers": {
"kazm": {
"url": "https://mcp.mellowmountainradio.com/mcp"
}
}
}
Cursor: .cursor/mcp.json ·
Windsurf: ~/.codeium/windsurf/mcp_config.json ·
Continue: MCP server block in config.json
Full docs and tool list: mcp.mellowmountainradio.com/docs
Smithery
Smithery is a public registry for MCP servers — it lets you discover and install them across many AI apps from one place. KAZM is listed there with all 45 tools. Click below to open the listing and connect from any supported AI client.
Open KAZM on Smithery →KAZM has been broadcasting from the mountain since 1974. The signal just got a lot farther.