Everything you need to create your first monitor.
Create a monitor, write your first assertion, and confirm alerting works in under five minutes.
Create your first monitor in the hosted product, write one assertion, and run the first check in a few minutes.
View setup guideTrigger a monitor from GitHub Actions or any CI runner and fail the job when the verdict fails.
View CI guideSee what gets stored, how your endpoint is called, and what stays visible in-product.
Review trust pageSee recent product changes explained in terms of what changed for you.
Read changelogHow PromptCanary notifies you
Every monitor has at least one alert destination. Alerts fire when a check fails and again when it recovers, so you always know when an incident opens and when it closes.
Available on every plan. Set one mail destination per monitor. PromptCanary sends a failure notice with the failing assertion and a diff link, plus a recovery notice when the check passes again.
Available on Pro and Team plans. Paste an incoming webhook URL into the monitor settings. Alerts post directly to the channel of your choice with the same failure and recovery context.
Available on the Team plan. Use outbound webhooks to route alert events to PagerDuty, incident tools, or your own pipeline. The full run result is included in the payload.
Turn a monitor into a deployment check
PromptCanary runs as a synchronous CI gate using a scoped workspace API key and a monitor id or slug — pass exits cleanly, fail returns a non-zero exit code so the workflow stops before rollout.
What to confirm before going live
- The first monitor can be created in under five minutes using a real endpoint and one high-signal assertion.
- Scheduler cadence, pause or resume controls, and recent cron decisions all look correct from the monitor detail page.
- Alert delivery outcomes are visible in-product and match the configured notification channels.
- Billing plan, retention window, and alert destinations match how you plan to use PromptCanary.