A Next.js app can pass every Lighthouse check on your machine and still fail a person on a mid-range Android with a flaky connection. Hydration cost, client-side data fetching and unhandled loading states are invisible locally.
The report combines what we can measure with what testers experienced, including the moments where the interface looked ready but was not interactive yet.
What you get: measured facts separated from human judgement.
Start an auditWhat we check
Hydration and interaction delay
We measure the gap between visible content and usable content — the interval where taps do nothing.
Loading and error states
Testers on slow networks report every place where a spinner never resolves or an error message says nothing.
Route-level metadata
Titles, descriptions, canonical URLs and indexability per route, including dynamic pages.
Onboarding completion
Testers sign up as new users and report the exact step where they got stuck or gave up.
Accessibility of custom components
Focus traps, keyboard navigation and contrast in hand-built UI, where the defaults do not save you.
Typical findings
- Client-side fetch on the landing page delaying the headline by seconds.
- Auth redirect loop when the session expires mid-flow.
- Dynamic routes sharing the same meta description.
- Modal that cannot be closed with the keyboard.
Questions
Can you test an app behind login?
Yes. Provide a test account and testers will run the authenticated flow.
Do you check the codebase?
No. The audit is black-box: what a visitor and a crawler can observe from outside.
Can findings be turned into tasks?
Yes, any finding can be handed off as a paid task for testers to verify after the fix.
See also
A roast is an automated audit plus reports from real testers, with evidence attached.