How we test

Every tool is signed up for or installed as a normal user would, on a paid tier where a free tier would distort the result, and put through the same three tasks. Nothing is scored from a vendor's marketing page.

The three tasks

1. A 60-second faceless short from a script. The same script each time. This tests text-to-video, voice, stock selection and pacing together.
2. Clip a 20-minute video. The same source recording each time. This tests whether a tool can find the moments worth cutting, not just cut.
3. Add captions and export. Tests caption accuracy, styling control, export resolution and what the free or entry tier actually lets out of the door.

What gets scored

• Output quality — would you publish it without re-editing?
• Speed — wall-clock time from input to finished export.
• Pricing transparency — can you tell what you will pay before signing up?
• Watermarking — what the tool puts on your video, and on which tier.
• Export limits — resolution, length and volume caps.

What this testing cannot tell you

Three honest limits. These tools ship changes constantly, so a score reflects the version tested on the date shown and nothing later. Three tasks are not a workload — a tool that suits our test script may not suit yours. And output quality is a judgement, not a measurement; where it is close, the review says so rather than inventing a decimal place to separate them.

Commercial relationships

Some tools pay commission on signups, and those pages carry a disclosure at the top. Scores are set from the task results before any commercial consideration, and paying vendors lose comparisons on this site where the results say they should. More on how the site makes money.