A fictional community repair-cafe invitation, 12 seconds, silent with captions.
0–4 seconds: Bring one object.
4–8 seconds: Explore what it needs.
8–12 seconds: Learn together.
Workflow A cuts between three views of the same generated illustration: the group, the lamp, and the wider group again. Workflow B uses a slow zoom across that illustration. Neither video depicts real people or documents an actual event. They compare editing choices, not competing generative video models.