Get started
Introduction
Send Xpensiv a link and you get back a fix list. Each finding has a screenshot, and you or your AI tool can work through it.
AI tools check that the code runs. They don't check whether the text is readable, whether the page fits a phone, or whether the buttons say what they do. Xpensiv checks those.
What we check
| Check | Question it answers | Returns |
|---|---|---|
| Readability | Can visitors read every word? | contrast ratio, text size, screenshot |
| Phone view | Does it fit a 390px screen? | overflow in px, screenshot |
| Wording | Do buttons and headings say what happens? | the text, a suggested rewrite |
| Layout | Are spacing and alignment consistent? | the element, a screenshot |
| Accessibility | Does it work with a screen reader and keyboard? | the rule, the element |
| Speed | Does it load quickly? | Lighthouse score |
Confidence
Every finding has a confidence score from 0 to 100%. Things we can measure, like contrast, score near 100%. Judgement calls, like wording, score lower. Filter by score to decide how much you want to see.
Using your fix list
Each finding says what's wrong, where it is and what it should be, so you can paste it into your AI tool. Then run the check again to see that it's gone.
Next steps
- Playground lessons: see what a check finds on a sample page. About a minute.
- Your dashboard: paste your own link and get your first fix list.
- API and MCP: run checks from Cursor, Claude Code or CI.