MCP Uptime

Know when your MCP server goes down — before your agents start misbehaving.

Uptime monitoring built specifically for Model Context Protocol servers. Pings every 60 seconds. Alerts by email when your endpoints fail, time out, or return malformed JSON-RPC.

Reserve early access — $29/mo Charged monthly, cancel anytime. Launches May 2026.

Why this exists

MCP servers fail silently. Your agent keeps asking for a tool, the server 500s or hangs, and you only notice when the downstream task has already failed three times. Generic uptime tools (Pingdom, UptimeRobot) check if the port is open — they don't know what a broken tools/list response looks like.

What you get

60-second health checks on every endpoint you register
JSON-RPC-aware probes — validates initialize, tools/list, resources/list actually return valid envelopes
Email alerts on failure (Slack / Discord webhooks coming post-launch)
Public status page per server (optional, for teams exposing MCP to partners)
Up to 10 endpoints on the early-access plan

Early access terms

Price

$29 / month

Launch

May 2026

Included

10 endpoints

You won't be charged until your dashboard is provisioned. Founding-member price locks for 12 months.

Who this is for

Anyone running an MCP server their agents depend on — internal tooling teams at companies rolling out Claude / GPT agent products, indie devs shipping public MCP servers, and ops folks inheriting an MCP fleet they didn't build.

Who this is not for

Hobby MCP servers you poke at from your own laptop. If downtime doesn't cost you anything, this isn't worth $29.

Reserve early access — $29/mo