Real MCP handshakes, not pings
We connect over Streamable HTTP or SSE and run the MCP initialize handshake, so you know agents can actually use your server. An HTTP 200 alone does not prove that.
Why join
Your AI agents are only as reliable as the MCP servers behind them
When a tool server silently breaks, agents fail in confusing ways. MCPServerStatus keeps watch around the clock and tells you exactly what broke, and when.
No credit card required.
Alerts before users notice
Get an email when a server goes down and another when it recovers, without alert storms while it flaps. Pro and Team add Slack and webhooks.
Latency you can see
Track handshake latency over time and spot slow deploys, cold starts and regional issues before they become outages.
Auth-aware checks
A 401 from a protected server means it is reachable, not down. Attach an encrypted auth header to check private endpoints end to end.
Incident history
Every outage is recorded with a start and end time and the error class (DNS, TLS, timeout, HTTP, protocol), so post-mortems take minutes.
Status badges
Show users your MCP server is healthy with a live SVG badge in your README or docs. It shows uptime too. Public status pages are coming soon.
Set up your first monitor in under a minute
Paste an MCP endpoint URL, pick a transport, done. Upgrade to Pro for 5-minute checks on 25 endpoints.
Start free