No green light nothing checks.
There is no public uptime page yet. Here is what is monitored, what happens when an engine fails, and how to tell us when something looks wrong.
Every alarm pages the people who build it.
Monitoring runs around the clock and posts to the team's Slack, with what broke and where to look.
The app is reachable
app.proofsource.co is checked over HTTPS from several regions every 30 seconds. About two minutes of failures pages the team.
Scans keep running
Alarms fire when a scan queue stalls, jobs start failing, a brand misses its scheduled checks, or an engine stops answering.
Answers are real
Alarms fire when answers come back empty or cut short at an unusual rate, so a quiet failure does not look like a lost mention.
An outage should not read as a lost mention.
Engines and data providers have bad hours. When one does, failed jobs are retried or deferred rather than closed, two sources switch to a fallback, and an alarm watches the rate of empty answers. A scan may finish later than usual.
| Source | If it fails |
|---|---|
| ChatGPT, Claude, Perplexity | Retried or deferred |
| Google AI Overviews | Second provider |
| Cited pages | Second reader |
About status and incidents
Is there a public uptime history?
Not yet. When there is, it will be read from the same monitoring described on this page, not typed in by hand.
What happens to my scan if an engine is down?
Failed jobs are retried or deferred rather than closed. Google AI Overviews and cited pages also have a second source, so one provider failing does not blank a scan.
How do I report a problem?
Email [email protected] with your brand and what you saw. A person replies within one working day.