Pick a depth. Each prompt opens in your AI pre-loaded with the lesson. Click a row to preview the prompt.
Pre-visualisation ('pre-viz') is the practice of generating rough, fast, disposable versions of your shots BEFORE you commit to the expensive high-fidelity versions. With AI tools, pre-viz is almost free — any model can produce a rough 5-second clip in under 2 minutes. The workflow: make the entire film in pre-viz quality first; edit it; iterate; THEN go back and re-generate the shots you kept at final quality. This is how professional AI filmmakers avoid the 'generate 200 beautiful clips that don't edit together' trap.
The entire 2-minute short assembled as 12 rough pre-viz clips in 40 minutes total. You edit these together in iMovie or DaVinci. You watch. You cut 2 scenes. You realise shot 7 needs to happen 20 seconds earlier. You re-arrange. You re-cut to music. NOW you know what the film is. THEN you regenerate the final-quality versions — and you've saved 80% of the budget that a 'generate-and-hope' workflow would have burned.
# Pre-viz workflow
Day 1 — Rough generation (4 hours)
For every shot in the shot list, generate ONE take at the cheapest/fastest setting.
Accept first-draft quality. Don't polish.
Label each clip: shot1A_previz.mp4, shot1B_previz.mp4, etc.
Day 1 — Rough edit (2 hours)
Drop all clips into DaVinci / Premiere / CapCut on a single timeline.
Arrange per the shot list order.
Watch it. Is the story working?
Cut any scene that doesn't earn its screen time.
Add placeholder music.
Day 2 — Refine (re-edit)
Tighten the rough edit. Audio timing. Cuts to beats.
Iterate until the pre-viz cut WORKS narratively.
Day 3–4 — Final generation (10–20 hours)
Only now: regenerate each surviving shot at the highest quality your tool offers.
Slot them into the locked timeline one at a time.
Don't re-cut — the edit was locked in pre-viz.
Day 5 — Finishing
Color grade. Sound design. Score.