Three readers
Three readers.
A human visual read of the rendered page at real viewports with a predicted attention map; a screen-reader read, linearised the way it is announced; a machine read of what crawlers can reach. Each ends in a short fix list with the measured fact behind it.
Free. No account. 3 reads an hour from one connection; a page read in the last 24 hours comes back as it was.
Three reads of one page.
- 01
What people see
- 5 viewports captured: three desktop, two phone. Up to 6 screens down the page.
- A saliency model predicts where the eye lands first.
- 30 AI agents per screen size read the 1440 desktop and 390 phone screenshots.
- 02
How a screen reader hears it
- The page linearised in the order a screen reader announces it.
- Headings, landmarks, links, images and form fields counted. axe-core structure rules.
- 6 AI agents briefed as blind or low-vision users listen to it.
- 03
What machines read
- robots.txt checked for 4 named crawlers.
- Words in the HTML as served, against words once JavaScript runs.
- JSON-LD, meta tags, canonical, llms.txt.
Now send 150 agents to the same page.
The live panel reads it the way your potential buyers would. Free account, no card.