Kinds of bots
Not all bots are the same, and a few of them matter a great deal to you.
- Search crawlers index your pages so people can find them. You want them, but they are not visitors.
- AI crawlers and assistants. Some collect text for training. Others fetch a page at the moment someone asks an assistant a question, which is a form of being read that you may care about.
- Monitors and tools. Uptime checks, link checkers and preview fetchers when a link is pasted in a chat.
- Scrapers and scripts. Copying content, probing for weak spots, or hammering a login form.
How bots get into the numbers
Most simple bots do not run JavaScript, so a script-based tool never sees them. The ones that do, such as headless browsers used for testing and some scrapers, can run the tracker like a person would. They then add pageviews, and often appear as a single-page session with no interaction, which pushes the bounce rate up. A sudden spike of visits from one country with almost no engagement is the typical sign.
How tools filter them
Filtering uses a few signals, none perfect on its own: the user agent string, lists of known data-centre address ranges, and behaviour such as no pointer movement. The result is always a trade-off. Filter too little and the numbers are inflated; filter too hard and real visitors, such as someone on an unusual browser, vanish. trckable filters on the server and tells you how many it dropped: the Visitors tile shows the number of bots filtered for the period, so a figure you doubt can be checked.
The robots a script cannot see
A tracker that runs in the browser cannot report a crawler that never opens one. That includes most AI assistants and search indexers. To see them you need the request itself, which your own server or CDN receives. trckable reports them from the server side: put its proxy in front of your site, or connect your CDN's logs, and the Links & bots section of Explore lists which assistants read your pages, which indexers visit, and which are collecting training data. The setup is covered in the features overview and the AI and search report.
What to do with it
Keep bots out of the numbers you use to decide things, such as signups per visitor. Read the crawler list separately: if an assistant fetches a page often, that page is probably one people ask about, and it is worth keeping accurate. A blocked crawler cannot be counted, so check robots.txt before concluding that no AI system reads you. The robots.txt tool helps you write one.
Try it in trckable
The Visitors tile says how many bots were filtered for the period, and Explore has a Links & bots section for the robots your server saw. See AI crawlers on your site
Related terms
Read more
The newsletter
Notes on measuring what matters.
Occasional emails from Albi: new posts, one chart worth reading, what changed in trckable.
We send one email to confirm your address, and nothing more until you confirm. Every newsletter has a link to leave. What we keep, and who sends it: privacy.