Heatpoints is a predictive attention platform that shows where visitors will look on a web page before it goes live. Paste any URL, staging link, mockup or screenshot and Heatpoints returns an attention heatmap, an attention score out of 100, and an AI-written audit with prioritized fixes, in about 30 seconds.
It runs UniSAL, a peer-reviewed saliency model (ECCV 2020) trained on real human eye-tracking data, so results are deterministic: the same design always produces the same score, usable for A/B decisions, CI checks and agent workflows.
Unlike session-recording and real-user heatmap tools, Heatpoints needs no tracking script, no cookie and no consent banner. The analysis runs server-side, so it works before you have any traffic and never touches your visitors. It also includes Words Map, a word-by-word reading-attention analysis of your copy that no other commercial tool offers, plus white-label PDF reports for agencies.
Ideal for CRO specialists, designers, marketers and founders who want to validate a page's visual hierarchy and messaging before launch.