Quick answer
Autocapture is automatic recording of user interactions — clicks, form submits, input changes, scroll depth — by the analytics script, without per-element tracking code.
What autocapture does
With manual tracking, someone decides in advance which actions matter and adds a track() call for each. Anything not planned is never recorded. With autocapture, the script listens to interactions on the page — which element was clicked, which form was submitted — and records them all. You decide later which ones are meaningful.
The big advantage is retroactivity: when someone asks "how many people clicked the new comparison table last month?", the data is already there even though nobody tagged it.
What autocapture cannot do
- Business meaning. A click on "Pay" is not a payment; the payment may still fail. Revenue, plan changes and server-side outcomes need explicit events.
- Stable names. Autocaptured clicks are identified by the element. When a redesign changes the page, the "same" button may look like a new element.
- Noise. Recording every interaction produces many events nobody needs; without a way to sort them, the useful ones get lost.
So autocapture works best alongside a small set of explicit business events, not instead of them.
Privacy in autocapture
Because autocapture records without per-element decisions, what it is allowed to record matters. A careful implementation records the element, not what was typed: no input values, no passwords, and a way to exclude whole areas of a page. Before turning it on, check what the tool collects by default, and whether element text is included.
Autocapture in TapCub
TapCub's autocapture records clicks on clickable elements, form submits, changes to selects, checkboxes and radio buttons, rage clicks (repeated clicks in one spot) and, optionally, scroll depth. Element text is off by default. Password fields, anything inside a data-tc-ignore area, selectors you block per site and the chat widget are never captured.
Captured interactions go to the AI event inbox, which clusters them and proposes a name for each pattern. You accept, merge or ignore candidates; accepted events are backfilled for 90 days and appear in event analytics like any tracked event. The same data feeds heatmaps, so click and scroll maps need no extra code.
Autocaptured candidates still need clear names before they become events; for rules and examples, see event naming conventions.
Common mistakes
- Relying on it for revenue. Send purchases and sign-ups from code or the server, where the outcome is known.
- Never reviewing candidates. Unreviewed autocapture is a pile of clicks. Name the few that matter.
- Turning on text capture blindly. Button labels are usually fine; text in other elements can contain personal data.
Related guides: How to Read a Heatmap and Event Tracking. See also track button clicks.
See also: User behavior tracking.
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Technical Engineer
Focuses on tracking code, SDKs, and API integrations, with practical guidance on implementation checks and troubleshooting.