Troubleshooting

Audits depend on the target site being reachable from our servers, on Lighthouse and Chrome completing within the configured timeout, and on the rest of the pipeline (screenshots, model pass) finishing successfully.

If something fails, the run detail page and the support inbox are the fastest ways to see what happened.

Confirm the URL is reachable in an incognito tab. Check for robots.txt blocks or anti-bot challenges. If the lab pass times out, the report indicates degraded lab data; re-run to confirm it is not a transient failure.

An audit gathers up to your plan's page cap (5 on free, 15 on paid) and only same-origin pages it can discover via the sitemap, robots.txt, and a shallow crawl. If pages are missing from the report, check that they are linked from the start page or listed in the sitemap, and that robots.txt does not block them. See Multi-page audits & sitewide scores.

First, confirm you are listening for an event that is actually delivered today: `report.created` and `run.completed`. The other event types can be selected but do not fire yet.

If deliveries are failing, check the deliveries log on the endpoint. The two usual causes are a signature mismatch (verify against the raw request body, not a re-serialized one) and an endpoint returning a non-2xx status. Fix the endpoint, then replay a past delivery. See Webhooks.

401 means the API key is missing, mistyped, or revoked: issue a fresh key and update where it is used. 402 means the organization is not on a paid plan, which API and MCP access require. See API keys and Plans & quotas.

Scheduling is paid-only, so confirm the plan is active. Scheduled runs also consume the monthly run quota; if the site has used its allotment for the month, the next scheduled run is skipped until the quota resets. Schedules are checked hourly, so a run starts near its time rather than exactly on it. See Scheduled audits.