Pasta Saved
Editor view

New scene

Each scene is one shot — one generation in Higgsfield or Seedance. Most generators work best with clips of 5–10 seconds.

How Pasta works

  1. Build the set. Add humans, boxes, a background and a camera. Give each one a clear name — the name is how the AI is told who is who.
  2. Move things. Select an object and drive it with the keyboard, a game controller, the mouse, the on-screen stick, or your own body through the webcam (Human → Track with camera).
  3. Record by timestamp. Press R: the scene plays and your movements are written into the timeline. Record each object in turn — the others play back while you perform. Re-record any part to punch in over it.
  4. One scene per shot. Add a new scene for the next shot; “Continue” carries the cast over to where they ended.
  5. Export for AI. You get a motion-reference video from the shot camera, the first frame as an image, and a timestamped prompt naming every object.

Keyboard

WASD / arrows
Move (relative to your view)
Q E
Turn
Shift / Alt
Run / slow
Z X
Down / up (camera, box)
T G
Tilt camera
R
Record / stop
Space
Play / pause
K
Keyframe here
Tab
Next object
V
Editor / shot view
Home
Back to start
Ctrl+Z
Undo (Shift to redo)
Del
Delete object

Game controller

Left stick
Move
Right stick
Turn · tilt camera
LT / RT
Down / up
A
Record / stop
B
Play / pause
X
Keyframe
Y
Switch view
LB / RB
Previous / next object
Start
Back to start

Pair the controller with your computer first, then press any button on it — browsers only reveal a controller after a button press.

Mouse & touch

Drag an object
Move it (records while recording)
Drag empty space
Orbit · right-drag pans · wheel zooms
Touch
On-screen stick and turn pads