Create and the models

Create is where you choose the model and write the prompt yourself. The composer is at the top of the page; the formulas and your avatars are below it.

How it works

1. Pick video or picture and a model.
2. Add a start frame, your products or avatars, or pictures, as the model takes them. Or use words only.
3. Write what you want in the prompt. Rewrite my prompt rewrites it for you, at no cost.
4. Set the length, shape, quality, sound and how many takes (1 to 4, each one charged).
5. Press Generate videos or Generate pictures. The credits it will use show beside it first.

The model menu says what each model takes. Anything it won't take stays on screen, greyed and marked Not sent, with the reason.

Video models

• Kling 3.0 by Kuaishou: 3 to 15 seconds, 720p or 1080p, sound if you want it (it costs more), start and end frames, and up to 3 elements.
• Kling 3.0 Omni: Kling 3.0 that also takes pictures. Frames, elements and pictures, up to 7 in all, at Kling 3.0's lengths and prices.
• Kling O1: frames, elements and pictures, up to 7 in all. 5 or 10 seconds, 720p or 1080p, silent.
• Seedance 2.0 by ByteDance: 4 to 15 seconds and up to 9 references. Standard goes up to 1080p; Fast and Mini cost less and go up to 720p.
• Seedance 2.5 by ByteDance: up to 30 seconds and up to 30 references, 480p or 720p.

Seedance's sound is on from the start and costs nothing extra. Seedance doesn't take people's faces, so avatars work with Kling and Nano Banana only.

Picture models

Nano Banana by Google, in three versions: Pro for the closest match to your product, 2 for lower cost, and 2 Lite, the cheapest, at 1K only. Up to 14 references, up to 4K.

Where results go

Finished videos and pictures appear on the page as they're made, and in your Library under AI media. Your plan sets how many generate at once: Creator 2, Pro 4, Business 8. More wait their turn.