APIzone — MCP server for API status

Is that API down? APIzone is independent, real-time status monitoring for 200+ popular APIs — Stripe, OpenAI, AWS, GitHub, Twilio and more. It runs its own probes from independent infrastructure, so status reflects what's actually observed, not what a vendor's status page reports.

This repository is the public home of APIzone's Model Context Protocol (MCP) server, so any AI agent — Claude, Cursor, or your own — can natively answer "is X down right now?" mid-task.

🔗 apizone.io · Status board · Incidents · API docs · MCP guide

Live examples (these badges update automatically):

Connect the MCP server

The server is hosted and stateless — no install, no build, no API key:

claude mcp add --transport http apizone https://apizone.io/api/mcp

Or point any MCP client at the Streamable-HTTP endpoint directly:

https://apizone.io/api/mcp

The server.json in this repo is the Model Context Protocol registry manifest.

Tools

All tools are read-only and require no authentication:

Agent discovery is also published at /llms.txt.

Related MCP server: GetMonitor MCP Server

Embed a status widget

Drop a live status widget into any page:

<div data-apizone="stripe,openai,aws-s3"></div>
<script src="https://apizone.io/widget.js" async></script>

Or a single badge: https://apizone.io/badge/<slug>.svg — generator at /badges, multi-API widget at /widget.

Public JSON API

No key required; 60 req/min per IP:

• GET /api/v1/status — all services with current status + latency

• GET /api/v1/status/<slug> — one API: 24h checks, 90-day uptime, incidents

• GET /api/v1/uptime/<slug> — uptime % and latency over 24h / 7d / 30d / 90d

Spec at /openapi.json; full reference at /api-docs.

How status is derived

• Synthetic checks. APIzone probes a public, unauthenticated endpoint of every monitored service. Any HTTP response under 500 counts as healthy (a 401 from api.stripe.com still proves the API is up); timeouts, network errors, and 5xx are failures.

• Status. Down = recent checks all failing. Degraded = some recent failures, or latency well above the service's own baseline.

• False-positive protection. Before opening an incident, the checker re-probes once; a transient blip that recovers is suppressed, so it never alerts subscribers.

• Independent. Status reflects APIzone's own probes and may differ from a vendor's official status page.

About this repository

This is the public MCP + product home for APIzone. The monitoring service itself is hosted at apizone.io; the application source is maintained privately. Issues and suggestions for the MCP server (missing APIs, tool ideas) are welcome here.