You started from a video on the wall. Change the words so they are yours, then continue: the voice and look are already chosen (you can change them in the next steps).
The helper is Google Gemini: your words are sent to Google for this. You can always change the result or undo it.
The pictures follow your text in order. Use ◀ ▶ to move one, ✕ to remove one.
1 · Template and tone
2 · Your words
The template fills this in as you type above; you can also edit it freely. One line = one subtitle. A blank line starts a new picture. [2] on its own line adds a 2-second pause (and a new picture).
3 · Voice
Not happy with how a line sounds? Change the take number for a new version.
Voice server settings
Not happy with how a line sounds? Change the take number for a new version. Each line is read separately, so keep lines short.
Read your text aloud, in one go. Your recording stays on your device.
Speech recognition runs in your browser: the first time it downloads a small model, your voice is never uploaded. Without it, timing is guessed from the pauses.
The placeholder hum is for testing timing, subtitles and video only.
4 · Pictures and clips
Pictures and clips follow the text in order. "Evenly" uses every picture and clip once, spread over the video length, switching at the nearest line. Clips play on a loop. Pack pictures are kept when a recipe is shared; your own uploads are not.
Your video
5 · Preview and export
More options
6 · Share your video
Please: no full names of private people, nothing about children, no links or requests for money. Reported videos are removed. You can remove yours at any time.
Save and load
A recipe holds the text, voice, settings and the style-pack pictures used. Your own uploaded pictures are left out. Whoever loads it starts from your video and changes the text.