A check on each preview link
Start it from CI, from Cursor or from the dashboard. Every report ends up in the same list.
For teams
Xpensiv for teams
If your team ships several apps a week with AI tools, nobody has time to look at every page on a phone. We check each release and send the list to whoever shipped it.
Talk to us about my teamWhat teams get
One bar for every app
The intern's weekend app and the main product get the same contrast, layout and wording checks.
Findings your AI tool can apply
Paste the list into the tool, ship the fix, and we check again to confirm it worked.
Your data
We don't train on your pages
Your pages and reports are not used to train any model.
Read-only
We open public pages in a sandboxed browser. We never write anything to your app.
Keys you control
Make or revoke API keys whenever you like. We store only a hash of each key.
Questions teams ask
Do you train on our pages or reports?
No. We use the pages you submit only to run the checks you ask for and show you the results.
Do you change our code?
No. We send findings. Your team or your AI tool makes the change, after someone on your side says yes.
Can we check private or staging apps?
Public URLs only for now, and that includes Vercel and Netlify preview links. If your staging is behind a login, tell us how it’s set up.
Can we run checks from our own tools?
Yes. Call the HTTP API from CI, or connect Cursor or Claude Code over MCP. You make API keys in your dashboard.
What data do you keep?
The link, the screenshots and the findings, so you can open a report again later. The Privacy Policy has the details.