Video
One clip from a line of text, a picture, or two keyframes.
The Video desk makes one clip at a time. It is the desk everything else is built on.
Four ways to start a take
- From words. Describe the shot. The model invents the whole frame.
- From a picture you give. Drop a still as the opening frame and the model films on from it — your product, your room, your photograph.
- From two keyframes. Give a start and an end frame and the brief describes the motion between them, not the frames themselves.
- From faces and products you saved. Attach passports and they ride the prompt, so the same face comes back take after take.
The desk picks the right one from what you give it: drop a frame and it switches to a picture-led take; tag a passport and it switches to a reference take.
Tagging what is in shot
Type @ in the brief for a face and # for a product or place. The menu lists what you have saved, and the tag reaches the model as a numbered reference — so "@priya holding #serum" becomes a shot with your Priya and your serum in it.
You can also drag pictures straight onto the brief. Each one gets its own tag.
Tools worth knowing
- Camera moves — push, orbit, track and the rest, each with a measured amplitude and speed.
- Techniques — shot grammar proven on this deployment rather than invented in the prompt.
- Templates — starter briefs in the model's own format.
- Match this style — point at a frame you like and the studio writes the look down and keeps it.
Keep exploring
- Passports — what makes a face survive between takes
- Storyboards — when one clip is not long enough