Two ways to monitor a site
| What | Free tools | Weekly site audit |
|---|---|---|
| Who runs it | Your agent, when you ask or on a schedule you set. | Our servers, in the background. |
| Site crawl | On your machine with the CLI. | Every week, with what is new, back or resolved since the crawl before. |
| Google index | 100 known URLs a run on the hosted MCP, more from the CLI, each checked live. | Up to the property's full 2,000 URL Inspection checks a day, top pages first. |
| Alerts | None. Your agent reports what it finds. | Email and webhook when a page drops out of the index or comes back. |
| History | Your agent keeps each result. | Kept on the server for each audited site. |
| Cost | Free, with no usage cap. | $50 a month per site. In beta. |
SEO alerts when a page leaves Google's index
Sites in the weekly site audit get alerts without asking. You get one when a page drops out of Google's index or comes back. A top page's drop is sent only after a second check agrees, so one odd answer from Google does not wake you up. Each run of a site's background checks that finds changes sends one alert for that site.
Each alert type has its own email and webhook switch, under Alerts on the Beta tab of your account page. Both are on by default, and the webhook starts once you save a URL. Email goes to your first linked Google account, and every email has a link that turns that type of email off. Webhooks are signed the Standard Webhooks way. After a timeout or a server error, a webhook is tried again up to 6 times over about a day.
On Monday morning a weekly summary covers every audited site: Google index checks, what the weekly crawl found and search clicks against the week before.
What the weekly crawl watches
- Internal links to errors, redirects and pages that could not be fetched.
- Missing or duplicate titles, missing descriptions and missing H1s.
- Pages that ask not to be indexed, and canonicals that point to another URL.
- Sitemap URLs that do not return 200.
- Thin pages under 100 words (a heuristic, not a Google rule).
- Pages where JavaScript changes the title, canonical or robots tag.
Each finding says whether it is new, back or resolved since the crawl before, so a release that broke something stands out from the problems you already knew about.
Start the weekly site audit
Pick a Search Console property on the Sites tab of your account page, or ask your agent:
Add example.com to the weekly site auditYour first site opens Stripe Checkout. Your agent can give you that link, but it cannot pay for you. The first crawl and index checks start once the plan is active, and you can cancel any time.
Monitor with the free tools instead
Run the index check once for a baseline and keep the full result. Run it again each week with the same number of pages and compare the two by page and reason. A new reason is the thing to look at. The seo command keeps no history here, so a scheduled agent task that saves each week's result works well.
seo index --site example.com --detail fullAfter a release, crawl the site and compare it with the crawl from before, and run seo compare on the release date once a week or two of Search Console data has come in. The report's traffic drop section looks for a sudden change in daily clicks that lasted, with the date it started.