Home / By task / AI text-to-video: what works and what is free
AI text-to-video: what works and what is free
A sentence can now become a video, but free access, network requirements and filing status differ a lot. These six tools are ordered by those three questions; every link goes to the tool's own site.
Tools you can use
Free allowance, clip length and resolution follow each site's current terms; we did not capture those numbers per site.
How to choose
- One 5–10 second shot: start with the ones that record a free tier (Vidu, Flova) before paying for a subscription.
- Chinese narration or a digital human: pick tools with digital-human or talking-head templates (Jimeng, Tongyi Wanxiang); a pure text-to-video model will fake lip sync poorly.
- Commercial use: read the site's terms; an ICP filing only covers the company, not likeness or brand rights — clear those yourself.
10-second shot template + four failure points
Template: subject + action + camera + light + duration. Example: “an orange cat peeks out of a cardboard box, slow push-in, late-afternoon window light, 10 seconds, realistic.” Write 9:16 for vertical, 16:9 for horizontal.
- No more than two camera moves: push, pull, pan and tilt together and the model drifts.
- State the aspect ratio first: 9:16 vertical, 16:9 horizontal.
- Do not let the model render text: on-screen words often come out garbled; add subtitles later.
- Hands and multi-person dialogue break first: start with one person and little hand movement.
Related pages
All AI video tools → · Short videos → · Digital humans → · All AI tools →