Product
Automation and API
Manage WatchCat from your code or your AI agent. An MCP server and a REST API cover every resource you manage in the app — behind scoped, owner-controlled API keys.
Hosted in the EU
GDPR-friendly by design
No US cloud in the core monitoring path
How it works
- Create an API key in the app, with a read or read and write scope
- Connect any MCP-capable assistant, or call the REST API directly
- Manage the same things you manage in the UI — monitors, channels, status pages, incidents
- Every key is scoped to your organization and respects your plan
terminal
# Connect an AI assistant over MCP
$ claude mcp add --transport http watchcat \
https://watchcat.io/mcp \
--header "Authorization: Bearer wc_…"
# …or create a monitor from your pipeline
$ curl -X POST https://watchcat.io/api/v1/service_monitors \
-H "Authorization: Bearer wc_…" \
-d '{"url":"https://example.com"}'
Technical flow
How it works in detail
Two surfaces, one set of keys, the same organization-scoped logic as the app.
Related docs
MCP server
- Connect Claude or any MCP-capable client to a single endpoint and let it work with your monitoring.
- Tools cover uptime monitors, cron monitors, groups, notification channels, integrations, status pages, status updates, incidents, and usage.
- A read-only key is offered only the read tools, so an assistant can answer questions without changing anything.
REST API
- A JSON API under /api/v1 for scripts, CI pipelines, and infrastructure-as-code.
- Standard create, read, update, and delete, plus pause and resume, cron token regeneration, and resolving incidents.
- A read-only key is limited to read requests; writes are refused.
API keys and scopes
- Keys are created by an organization owner and shown once, then stored only as a hash.
- Each key is read-only or read and write — pick the least access a task needs.
- Revoke a key at any time and the clients using it stop immediately.
Safe by default
- Every request is scoped to the key's organization; there is no path to another tenant's data.
- Plan limits are enforced exactly as they are in the app.
- Secrets such as webhook URLs and bot tokens are never returned in responses.
Why it matters
- Define monitors as part of your deploy pipeline instead of clicking through the UI
- Let an AI assistant triage incidents, resolve them, and post status page updates
- One set of keys, two ways in — MCP for agents, REST for scripts and CI
- Read-only keys make it safe to give an assistant visibility without write access
What you can manage
Both the API and the assistant reach every resource you manage in the app.
- Uptime monitors
- Cron monitors
- Monitor groups
- Notification channels
- Integrations
- Status pages and updates
- Incidents
- Plan usage
Start monitoring for free
5 monitors free · no credit card required