For maintainers
MCP servers break silently. Monitor tells you first.
A dependency bump, an API change, a bad release, and your server stops speaking MCP without a single error reaching you. The usual way maintainers find out is an angry GitHub issue days later. Monitor runs a real handshake against up to 5 of your servers every day and emails you the moment something changes.
Prices exclude tax; applicable tax is added at checkout.
A daily live handshake
Every monitored server is spawned and taken through a real MCP session daily (initialize, tools/list), exactly like a client would. Not a ping, the actual protocol.
Tool-drift alerts
A tool disappeared, or a new release changed the list. Every agent integration depending on your server cares, and nobody else tells them. You get one email naming exactly what changed.
Alerts only on change
Stopped responding, recovered, or drifted. That is the whole inbox. Steady states never email you. The product is the alert, not the noise.
A public status page
Every monitored server gets mcpvault.io/status/your-server: a 90 day uptime strip, last-check facts and an incident log. Link it from your README so users can see the server is watched and working.
Pro included
Monitor includes the full Pro membership on every server you own: badge-free verification, the do-follow link, pink-highlighted listings. One subscription.
How it works
- Subscribe (yearly $59 or monthly $7; prices exclude tax, added at checkout).
- Pick up to 5 of your claimed servers on the dashboard.
- The first daily check sets the baseline; from then on you are emailed only when the server fails, recovers, or its tool list changes.
- Each monitored server gets a public status page at /status/your-server: uptime history, incidents, last-check facts. Link it from your README.
- Cancel any time from your account. Monitoring runs until the end of the paid period.
Not sure your server passes at all? Run the free tester first.