Integrations
A crawl can only see your site
A crawler cannot see who came, what they typed, or what they did next. Four connections add that half of the picture.
Who came to read it
Every crawler that reached your site and what it got back. Googlebot, Bingbot, GPTBot, ClaudeBot, PerplexityBot - by day, and by response code.
Cloudflare in detail →What they searched for
The queries people actually typed, the pages that answered them, and where those pages sat. Google's own numbers rather than an estimate of them.
Google data in detail →What they did next
Engagement and conversions per landing page, joined to the findings on that page. The difference between a slow page and a slow page that loses people.
Google data in detail →What the other index holds
How many of your pages Bing has, and what its crawler hit getting them - the index behind DuckDuckGo, Yahoo, Ecosia and Copilot.
Bing in detail →Your crawl budget, measured instead of guessed
Every technical SEO tool talks about crawl budget from log files nobody exports. Cloudflare already has them, so the question stops being theoretical: how often does Googlebot actually come, and does it get a page or a 404?
On the first site we connected, Googlebot averaged 86 requests a day while Applebot averaged 3,509 - forty-one times more attention from a crawler that does not decide whether you are indexed.
The crawler that decides whether this site is indexed came least of all.
Whether the AI crawlers can read you at all
The AI visibility page tells you whether an assistant cites you. This tells you whether its crawler could get in - and the two disagree more often than anybody expects, because a robots.txt rule set years ago is still being enforced.
On that same site, Google-Extended was refused 40.6% of the time, GPTBot 22.6% and PerplexityBot 21.6%, while Googlebot was refused 2%. That is not a ranking problem; it is a door.
Same site, same week. The search crawler gets in; three of the AI crawlers largely do not.
Read-only, and the narrowest scope each vendor offers
Cloudflare is a token you create yourself with one permission, on one zone. The two Google sources go through Google's own consent screen, read-only. Every credential is encrypted before it is stored and none is ever shown back to you.
Disconnecting takes the data with it: the Cloudflare rows are deleted, and revoking a Google source removes what it filled.
What each one unlocks
Crawler traffic by day
Which bots arrived, how often, and what they got - with the error rate per crawler, because arriving and being refused is not the same as reading.
Queries and landing pages
What people typed, which page answered, and where it sat. Joined to the findings on that page, so a fix has an audience attached.
Behaviour after the click
Engagement per landing page next to that page's technical findings - which is what turns a list of fixes into an order of fixes.
One page per site
Connections live under the site they belong to, not in a global settings screen, so it is always clear which site a source is feeding.
And one that works the other way
The three above bring data in. This one puts the result where the work happens.
Scan when you deploy
A deploy key starts one scan of one site and can read nothing - so it lives in your CI secrets without carrying the rest of your account with it. One POST after your build, from GitHub Actions, Vercel, Netlify or anything else that runs a command.
A scan already running is returned rather than duplicated, and a daily cap bounds what a key pasted into a job that runs on every commit can cost.
How to call it →Questions
- No. The audit, the score and the fixes work on their own. These add data the crawl cannot reach - each one is useful on its own and none is a prerequisite.
- Read-only, and the narrowest scope each vendor offers. Cloudflare is a token you create with one permission - Zone · Analytics · Read - on one zone. Bing is a read-only API key from your own Webmaster Tools account. Search Console and Analytics are Google's own consent screen, read-only. Every credential is encrypted before it is stored and none is ever shown back to you.
- Bing is the index behind DuckDuckGo, Yahoo, Ecosia and Copilot, so it decides more than its own search share. It also reports how many of your pages it holds and what its crawler ran into - 4xx, 5xx, robots blocks - which Search Console has no equivalent screen for.
- Eight days on a free Cloudflare plan and more on paid ones. We ask one day at a time and stop where your zone stops, so you get whatever your plan keeps without configuring anything.
- Yes, from the site's Integrations page. Disconnecting Cloudflare deletes the traffic we pulled; disconnecting a Google source revokes our access at Google and removes the rows it filled.
Start free
One site and 100 credits - up to 200 pages a scan, about 5 scans, and the whole report. No card. Paid plans add more sites, bigger crawls and the AI rewrites.