Welcome
snorklee shows you who comes to your site: the AI — assistants, robots, automated systems — that read it, fetch one of your pages to answer someone, and then send you visitors. Alongside them, we keep your human visitors, so you always have something to compare against. For every visit, we tell you what we could check: either its address matches the robot maker's official ranges, so it really is that robot, or it just gives us its name, with no proof. Everything is observed on your server, nothing is guessed, with no measurement cookie and no advertising identifier. This help does four things: install the Snorklee code, read what the AI are doing, compare it with your human traffic, and keep a record for your compliance.
Where to start
- Copy the Snorklee code from the Integration tab.
- Paste it once into your site's
<head>, or follow your platform's guide. - Install the server module that catches robots: robots don't run the JavaScript on your pages, so without this module you won't see a single AI go by.
- Open Verify installation, then the Agents tab to watch the first visits arrive.
- Come back to Actions: as soon as we know enough, Snorklee puts what needs deciding there, with the evidence behind it and how we'll measure whether your decision paid off.
What snorklee measures
- Who comes to your site, by type: AI robots that crawl, AI search-engine robots, assistants that read a page while answering someone, visitors sent by an AI, and humans.
- How sure we are about each visit — always stated separately: "400 GPTBot visits" and "400 visits from something that presents itself as GPTBot" don't mean the same thing.
- The path the AI take: the pages they explore, the ones they read live to answer, and the ones that end up sending you visitors.
- The comparison between your humans and the AI, in the same unit and over the same period.
- For your human visitors: pageviews, where they come from, countries, devices, errors and broad usage trends — all grouped together, never tracked person by person.
- What your robots.txt file asks for, compared with what robots actually came to fetch. A robots.txt is a "no entry" sign, not a lock: a robot can ignore it.
What snorklee does not see
- AI that drive a real browser (for example with Playwright, or a "computer-use" agent): the Snorklee code doesn't start in an automated browser. We still name them in the AI traffic tab, marked "we can't count these" — never a zero, which would suggest there are none.
- Programs that fetch your pages without a browser: dropped on arrival, never stored.
- Stealth AI that pass themselves off as an ordinary browser: impossible to tell apart from a human today, so they're counted among the humans.
- No measurement cookie, no identifier that would follow a visitor from one visit to the next, no session replay, no individual profile.
What to keep in mind legally
snorklee helps you keep a record of a deliberately minimal measurement. It does not replace your own compliance: your site, your legal basis, the information you give visitors and your other tools remain your responsibility.