How we test

Every verdict on this site comes from the same process. No press accounts, no sponsored placements, no scores for tools we haven't used.

The rules

• We pay or use the real free tier. Whatever a normal customer gets, that's what we test. No vendor demos, no reviewer accounts.
• Identical test projects. Head-to-head tools get the same script, the same brief, the same deadline. For AI video that means one 30-second script through every generator.
• We measure cleanup, not demos. The number that matters is how much fixing a draft needs before it's publishable - not how good the vendor's showcase looks.
• Pricing is captured from the real checkout. When we list a price, we saw it on the vendor's own signup or checkout screen, and we say when.
• Watermarks and limits get named. Free-tier catches are part of the verdict, not a footnote.
• Verdicts publish with proof. Scores land alongside the video showing the actual test, screenshots from our projects, and the plan we used.

What "review in progress" means

Pages for tools still in the lab are labeled "review in progress" or "verdict pending." They contain vendor-stated features (marked as such), facts we have already verified ourselves (marked with a date), and what we're testing next. They contain no scores.

Updates

AI tools change monthly. When a tool ships a major update or a price change, the affected pages get re-tested and the change is noted on the page.