A daily check of your site's AI visibility
Robots.txt is only half of what decides whether an AI search crawler reaches your site. A firewall or bot rule can refuse the crawler behind ChatGPT search or Perplexity while your robots.txt welcomes it, and a robots.txt check cannot see that. In September I found 56 of the top 5,000 sites doing exactly that, most of them with a block list that turns away the AI training bots too (the data). Monitoring runs that server test on your site every day.
The free checker tells you where your site stands today. Monitoring keeps checking after you close the tab: a re-check of your site every day, with every result and change kept on a private status page that also has an Atom feed. You also get an email when a watched status changes. $19/mo after a 7-day free trial, cancel anytime. See a live example of the status page.
How you follow it
Each subscription gets a private status page for its site: the current status of every check, a dated entry for each change, and the day-by-day history of the last 30 checks. The address is a long random link that only the subscriber has, and the same address ending in .atom is a feed for any feed reader. You also get an email at the address you subscribed with: a welcome note with your private link, a first report after the first check, and a short plain-text note whenever a watched status changes. The page and the feed keep the full record.
What the daily check covers
- Crawler access in robots.txt - whether the AI search crawlers behind ChatGPT and Perplexity (OAI-SearchBot, PerplexityBot) are allowed.
- A test at the server as well as robots.txt - a firewall or bot rule can refuse a crawler that your robots.txt welcomes, and nothing in robots.txt shows it. The check requests your homepage as OAI-SearchBot and as PerplexityBot, and runs two control requests before it calls anything a refusal.
- Readable text - whether your homepage ships readable text without JavaScript.
- llms.txt - whether your llms.txt is there.
- A private status page - the current status of each check, every change with its date, and the latest 30 checks. It is at a random address that appears on the page you land on after paying, so bookmark it. It is also in every email we send you. Anyone you give the address to can read the page.
- An Atom feed at the same address - add it to a feed reader and each change arrives as a new entry.
- Email - a welcome note with your private link, a first report after the first check, then a short plain-text note when a watched status changes. Status only, never promotion.
- A public result page - your site's current status stays visible to anyone at /r/yoursite.com, monitored or not.
The honest limits
- One site per subscription.
- Checks run once a day, not in real time.
- The first check runs at the next scheduled run (13:00, 16:00, 19:00 or 22:00 UTC) after you subscribe. Until then your status page says no check has read your site yet.
- The check runs from a scheduled job at 13:00 UTC, with backup runs at 16:00, 19:00 and 22:00 UTC that only do anything when no run in the last 12 hours produced a check. If a day still has no check, the list on your status page shows a gap for it.
- Email is plain text, status only, and goes to the address you subscribed with. A message can land in a spam folder or be delayed, so the status page is the record to trust. There is no push message.
- When the server test cannot tell, because the request timed out or the site's firewall also turns away test requests that name no crawler, the status page shows the robots.txt result alone and does not mark the server as refusing.
- The server test sends the crawler's name from Cloudflare's network, so it catches rules that match the user agent. A rule that lets in only the crawler's real IP addresses will not show up.
Already subscribed?
Manage or cancel anytime in the subscription portal (enter the email you subscribed with).
Questions
What counts as a change? A real status flip in the daily check: an AI search crawler newly blocked or unblocked in your robots.txt, your server starting or stopping refusing one of them, your homepage's readable-text status changing, or your llms.txt appearing or disappearing. A change shows up as a dated entry on your status page and in the feed.
What happens right after checkout? Stripe sends you back to your site's result page, which shows your private status link. Open it and bookmark it. The status page says no check has read your site yet until the next scheduled run, then it fills in. Your card is not charged for 7 days. You also get a welcome email with the same link, and a first report by email after the first check.
How do I cancel? From the subscription portal, anytime, including during the 7-day trial with no charge at all. It is Stripe's standard self-serve portal; you do not have to email anyone, though reese@lastminutedealshq.com works too.
Why pay when the checker is free? The check is free and stays free. The 7-day trial is also free. The $19/mo after that pays for the standing daily re-check and the record it keeps, the part that keeps watching after you close the tab.
Free, with or without a subscription
The free checker shows where your site stands right now, and every site has a public result page at /r/yoursite.com that runs the same check. /changes lists the sites that changed since the September census. The next census re-run is planned for early October, and its results go on /data.
For agencies
If you fix AI search visibility for clients, the list page has the sites whose server refuses ChatGPT's or Perplexity's search crawler while their robots.txt allows it, with the test dates and a free 25-row sample.
Not ready to pay?
Leave your site and an address and I will count you for the October census re-run. Nothing is emailed; the figures go on /data.
Not sure yet? Run the free check first.