create-monitorLABSCO SUMMARY
Monitoring you can set up and operate without opening the dashboard.
The coverage runs well past creating monitors: maintenance windows suppress alerts around planned work, status pages and announcements are how you tell users what is happening, and incident comments keep the write-up where the incident is. The quick-monitor-setup path is the unusual one — a free HTTPS monitor from an email address alone, confirmed by clicking a link — which makes "start watching this URL" something an agent can do for a person who has never signed up.
What it is
UptimeRobot's own hosted server at mcp.uptimerobot.com, reached through mcp-remote with an OAuth browser flow. It covers monitors and monitor groups, public status pages and their announcements, maintenance windows, incidents and their comments, plus alert integrations and tags.
What you get
- Monitors created in any of its types — HTTP, KEYWORD, PING, PORT, HEARTBEAT, DNS, API, UDP — then renamed, re-pointed, re-intervalled, paused or resumed
- Monitors listed with search and state filters, one monitor's full config and current state, and aggregated up, down and paused counts with uptime percentage
- Bucketed response-time series, for looking at a slowdown rather than only an outage
- Public status pages built from monitors, groups or tags, with domain, sort order, visibility and publish status
- Status-page announcements posted as info, maintenance or issue, then edited, pinned and unpinned
- Maintenance windows scheduled one-off or recurring, so alerts stay quiet through planned work
- Incidents listed across monitors by time range, and one incident opened up to its checker locations, logs and traceroute
- Comments added to an incident and edited, keeping the write-up attached to the incident itself
- The alert-contact integrations available on the account, and the account's tags
- A free HTTPS monitor created from someone's email address alone — no account, no key — confirmed by clicking a link in the activation mail
Requirements
An UptimeRobot account. The default route is npx -y mcp-remote@latest https://mcp.uptimerobot.com/mcp, which opens a browser once for OAuth and caches the token — no API key to paste. Where no browser can run, connect over plain HTTP to the same URL with a Main API key from the dashboard's Integrations & API page, sent as an Authorization: Bearer header. Monitor types, minimum intervals and monitor counts follow your plan and the server enforces them, returning -28002 when a call would go past what the plan allows; monitor groups, maintenance windows and incident comments are themselves plan-gated. The repository also ships Claude Code, Cursor and Codex plugin manifests that register the server for you.
Setup effort
One command, then sign in — claude mcp add uptimerobot -- npx -y mcp-remote@latest https://mcp.uptimerobot.com/mcp, then approve the OAuth sign-in
USE CASES
01
Putting a new service under watch
Create the monitor with its type and interval, attach the alert contacts, then read the config back to confirm it is reporting.
02
Keeping the alerts quiet during a deploy
Schedule a maintenance window over the change, one-off or recurring, and extend it if the work overruns.
03
Working out what happened last night
List the incidents in that window, then open one for its checker locations, logs and traceroute.
04
Telling users about it while it is happening
Post the announcement to the status page as an issue, pin it while it matters, and edit it as things resolve.
05
Standing up a status page
Build the page from monitors, groups or tags, set its domain and sort order, then publish it.
06
Checking whether it is getting slower, not just failing
Pull the bucketed response-time series next to the uptime percentage.
07
Writing the incident up where people will find it
Add the comment to the incident itself rather than to a document nobody opens.
08
Tidying an account that has sprawled
List the monitors with filters, group them, and check which alert integrations are actually in place.
32 tools
Listed in project docs. Your client may show a different list.
update-monitorupdate-monitor-statuslist-monitorsget-monitor-detailsget-monitor-statsget-response-timescreate-monitor-groupupdate-monitor-groupget-monitor-grouplist-monitor-groupscreate-pspupdate-pspget-psplist-pspscreate-psp-announcementupdate-psp-announcementpin-psp-announcementunpin-psp-announcementget-psp-announcementlist-psp-announcementscreate-maintenance-windowupdate-maintenance-windowget-maintenance-windowlist-maintenance-windowslist-incidentsget-incident-detailscreate-incident-commentupdate-incident-commentlist-incident-commentslist-integrationslist-tagsBEFORE YOU INSTALL
What to know first
Tools you get
32 tools named in the project's documentation.
Requires
npx on PATH
INSTALLATION
Connect your client
Server setup
claude mcp add
TERMINAL
claude mcp add uptimerobot -- npx -y mcp-remote@latest https://mcp.uptimerobot.com/mcpHosted endpoint
ENDPOINT URL
https://mcp.uptimerobot.com/mcpAdd this hosted endpoint to your MCP client, then sign in.
MCP Clients
Claude Code, Cursor, Other MCP clients
JSON
{
"mcpServers": {
"uptimerobot": {
"command": "npx",
"args": [
"-y",
"mcp-remote@latest",
"https://mcp.uptimerobot.com/mcp"
]
}
}
}Claude Code — plugin
CLAUDE CODE — PLUGIN · TERMINAL
git clone https://github.com/uptimerobot/ai.git
claude --plugin-dir ./aiCodex
CODEX · TERMINAL
git clone https://github.com/uptimerobot/ai.git
cd ai
codex plugin marketplace add ./Any MCP client — API key
ANY MCP CLIENT — API KEY · JSON
{
"mcpServers": {
"uptimerobot": {
"type": "http",
"url": "https://mcp.uptimerobot.com/mcp",
"headers": {
"Authorization": "Bearer <YOUR_API_KEY>"
}
}
}
}
