The score changes when nothing changed.
When a model invents the number, the same page scores differently on every run. You can never tell if your work helped.

Paste a URL. UXLens checks your site on desktop, tablet, and mobile with real WCAG tests, gives you a measured score, and shows you exactly what to fix.
No signup for your first audit. Results in about 90 seconds.
An opinion on look and feel. Never changes the score.
Example audit. Real reports name the exact element and give you a fix to copy.
Scores that drift when nothing changed, and fixes that never show up in the number. UXLens was rebuilt to remove both.
When a model invents the number, the same page scores differently on every run. You can never tell if your work helped.
Reports mix opinions with measurements, so you work through a list of suggestions and nothing improves.
UXLens scores only what it can measure. Run it twice, get the same number. Fix a scored finding, and the number goes up.
See what gets measuredEvery scored check runs on your rendered page at three screen sizes. If a developer can fix it, it can move your score. If it is a matter of taste, it stays advisory.
No model picks the number. Identical pages get identical scores.
Real color math on every text element, including semi transparent layers.
Every tappable element on mobile, measured against the 44px minimum.
Desktop, tablet, and mobile are rendered and scored on their own.
You always know which findings move the number and which are advice.
Apply the fix text as written, audit again, and watch the score climb.
axe-core, contrast, alt text, keyboard focus, form errors
Tap targets, sideways scroll, viewport, 16px text, line height
One H1, heading order, reading level, line length
Broken links, image loading, page title, contact details
Four steps, and the last one is the proof.
Any public page. No install, and no signup for the first audit.
Up to 12 pages at three screen sizes, with axe-core WCAG checks and real screenshots.
Each finding names the element and gives a fix for your code or your AI.
Fixed findings drop off and your Measured UX Score goes up.
You know how to build. UX standards are a different job. Run an audit before launch and catch the contrast failures, tiny tap targets, and missing states your users would have found first.
Audit my siteWhat you get
Your agent builds a page, calls UXLens, applies the fixes it gets back, and audits again until the score stops moving.
curl -X POST https://uxlens.io/api/audit \
-H "Content-Type: application/json" \
-H "X-API-Key: $UXLENS_API_KEY" \
-d '{"url": "https://my-site.com"}'{
"score": 7.2,
"scored_issues": 6,
"advisory_issues": 11,
"fix_prompt": "Paste into your agent...",
"report": "https://uxlens.io/r/abc123"
}A curated library of UX and accessibility standards, with a plain explanation of every check UXLens runs and where each rule comes from.
API reference, authentication, response schemas, and setup guides for agents and CI.