MCP server: ask your AI assistant what’s down
You already ask an AI assistant to help debug. With the StatusEagle MCP server, it can check your monitors itself: what’s failing, since when, from which locations, and what else broke at the same time.
An MCP server (Model Context Protocol server) exposes a product’s data and actions to AI assistants in a standard way. The StatusEagle MCP server lets assistants such as Claude, ChatGPT or Cursor read your monitors, status and incidents when you ask.
What you could ask
How it will work
{ "mcpServers": { "statuseagle": { "command": "npx", "args": ["-y", "PACKAGE_NAME"], "env": { "STATUSEAGLE_API_KEY": "YOUR_READ_ONLY_KEY" } } } }Configuration shape for MCP clients. The package name and any hosted endpoint will be published at launch.
{
"mcpServers": {
"statuseagle": {
"command": "npx",
"args": ["-y", "PACKAGE_NAME"],
"env": { "STATUSEAGLE_API_KEY": "YOUR_READ_ONLY_KEY" }
}
}
}Read-only first, on purpose
An assistant that can read your monitoring is useful. One that can pause monitors or post to your status page needs guard rails. The first version is read-only. Any future write actions, such as acknowledging an incident, will need a separately scoped key and your assistant’s confirmation step.
What you can do today
The REST API already exposes monitors, a status overview and ongoing incidents. If you build your own tools for an assistant, those endpoints are what the MCP server will wrap.
Related monitoring
Pair these monitoring checks to ensure complete stack coverage.
Frequently asked questions
Still deciding? Start monitoring free and verify your first service in ~2 minutes.
A server that implements the Model Context Protocol, an open standard that lets AI assistants call tools and read data from other applications.
Your first monitor takes about two minutes.
Free plan: 10 monitors, 3 locations, 1 status page. No credit card.