See every monitor live
One screen. Test regions, AI-diagnosed incidents, latency charts, and a timeline your agent can read.
AI-native monitoring
You stop finding out from your customers.
We check your endpoints from up to 12 regions, every sweep. When something breaks, the diagnosis is already written by the time your phone buzzes.
Runs 6 live checks on any public URL. No account, no card.
- 10 checks per sweep
- 12 test regions
- Free tier, 3 monitors
A monitor that explains itself.
Most tools tell you something is down and leave the diagnosis to you at 3am. UpMonitor correlates every region, every checker and the deploy timeline, then hands you a written finding and a prompt you can paste straight into your own agent.
Every number on this page is produced by the running system. We do not publish figures we cannot show you the source of.
Check → Diagnose → Fix
Three primitives, one loop. Every monitored site flows through them every minute.
- 01
Check
10 checks from 12 regions, every sweep. SSL, DNS, DNSSEC, security headers, liveness, page weight, Core Web Vitals and more - the full list is in the rail at the top of this page.
- 02
Diagnose
Vertex AI reads the payload, identifies the root cause, drafts a fix prompt and exposes it via the upmonitor:// protocol.
- 03
Fix
MCP server hands the incident to Claude Code, Cursor, Continue, or Cline. The agent fixes it. You wake up to a closed ticket.
Your SRE Diagnostic Console — in your browser
Every monitor, every check, every incident. Status pills update live. The dashboard mirrors what your AI agent reads through MCP.
SSL Renewal Watch
90 / 60 / 30 / 14-day expiry alerts. Chain-of-trust validation per region.
DNSSEC Chain
Full DS / DNSKEY / RRSIG validation. Catches drift before the .io zone breaks resolution.
MCP Server
Built-in. Your agent reads incidents directly — no glue code, no scraping dashboards.
Lighthouse+
Core Web Vitals enriched with font self-hosting, TTFB DNA, and AI-readable JSON output.
$ upmonitor monitor add upmonitor.io → reading config from .upmonitor.yml ✓ probe deployed to us-central1 ✓ probe deployed to europe-west1 ✓ probe deployed to asia-northeast1 ⚡ dispatched to MCP server
When something breaks, you see the triage first
The Diagnostic Inbox is your single triage surface. Filter by check or by monitor. Acknowledge what's owned. Resolve what's fixed. Hand the AI a pre-built prompt for what's still smoking.
Open the Inbox →
Drop into your agent's surface
The UpMonitor CLI ships on npm. The Model Context Protocol server runs locally. Same probes, same payload — from your CI, your IDE, or your agent.
- npm @upmonitor/cli + @upmonitor/mcp-server
- MCP-ready for Claude Code, Cursor, Continue, Cline
- upmonitor:// protocol for one-click triage
From alert to fix — without a human in the loop
When SSL fails or DNS drifts, UpMonitor sends the full test results to your agent. Your AI reads, decides, fixes.
One platform. Three tiers.
Start free with 3 monitors. Move to PRO for 1-minute checks and MCP. Agency unlocks 100 monitors and team seats.
Free
One or two services, watched properly. Not a trial.
1 of 12 regions
$0/mo
Start free 3 Monitors 5min Check Interval 1 Region 30-Day History 5 AI Diagnoses / month Email Alerts Standard Lighthouse
Most Popular
PRO
A real fleet, six regions, AI diagnosis included.
6 of 12 regions
$9/mo
Choose PRO → 25 Monitors 1min Check Interval 6 Regions 90-Day History 50 AI Diagnoses / month MCP server + CLI Lighthouse+ AI, webhooks
Agency
Multi-tenant, unlimited AI triage, team seats.
All 12 global regions
$32/mo
Choose Agency → 100 Monitors 1min Check Interval 12+ Regions 365-Day History Unlimited AI Diagnoses Agentic webhooks (MCP URIs) Public status pages & seats
Free Website Monitoring Tools
Run any check without an account — SSL, DNS, HTTP, security headers, Lighthouse+, and more. Each tool is scriptable from CI and returns JSON your AI agent can parse.
Frequently Asked Questions
MCP is an open standard that lets AI agents like Claude and Gemini securely access data and tools. UpMonitor ships an MCP server so your agents pull full diagnostic context directly during an incident - no manual copy-paste from dashboards.
An Agentic Webhook delivers a secure Model Context Protocol (MCP) URI to your AI agent, pointing it at the exact logs, configs, and history needed to fix the incident. Standard webhooks only say "System Down"; UpMonitor hands the agent everything required to remediate.
Yes. UpMonitor is free forever for 3 monitors at 5-minute intervals - ideal for personal sites, portfolios, and indie projects.
Yes. The UpMonitor Shadow Monitor API is stateless and built for pre-deploy audits. Run a full SSL, DNS, and security headers check on every commit - one curl command, zero setup - and fail the build before bad config hits production.
Every UpMonitor alert includes a `upmonitor://monitor/<id>` link. Run `upmonitor register` once on your machine, then any click on that link from Slack, email, or an agent UI opens your local terminal loaded with the full diagnostic context - no copy-pasting IDs, no hunting for logs. Automated playbook execution is on the roadmap.
Free accounts run 5-minute checks. PRO and Agency both run 1-minute checks. PRO covers 6 global regions; Agency covers 12+ regions and adds Agentic Webhooks plus public status pages.
UpMonitor sends an instant email alert. PRO adds push notifications (PWA), Slack, and custom webhooks. With the MCP server enabled, your AI agent also receives a structured Agentic Webhook with full diagnostic context - and can start the repair autonomously.
Lighthouse+ is an enriched performance engine that adds proprietary audits - font self-hosting, TTFB DNA, and more - on top of standard Lighthouse data. The output is formatted for AI agents: they identify, apply, and verify performance fixes automatically via the CLI and Shadow API.
Every degraded monitor surfaces an Active Incident panel at the top of its details page - showing the root cause, three remediation steps, and a copy-pasteable prompt that hands the entire incident off to your AI agent (Claude Code, Cursor, Claude Desktop, or any MCP-compatible client). Two clicks from "site down" to "agent has the fix." Read the full lifecycle in the <a href="/docs/active-incident">Active Incident docs</a>.