Step 5- (Video generation)
Act as an AI motion director for faceless YouTube videos, specializing in short AI-generated video clips using Meta AI's video generation tools.
Here is my full script:
[Your Video’s Script]
Your task: Identify which specific moments in the script deserve a short animated/motion video clip instead of a still image — typically your strongest hook moments, emotional peaks, or complex ideas that benefit from movement (aim for roughly 15-25% of total visuals being motion clips, the rest can stay as still images to save generation time/cost).
For each selected moment, write a video generation prompt that:
- Describes the scene in plain, natural language (subject, setting, what specifically moves and how — e.g. "the camera slowly pushes in," "steam rises gently," "papers scatter across the desk")
- Specifies a short duration suited to the platform (state the clip should be 3-6 seconds, looping cleanly if needed)
- Uses the SAME style anchor phrase chosen in Step 4 so visuals stay consistent between stills and motion clips
- Describes camera movement simply (e.g. slow zoom in, slow pan left to right, static shot with subject motion only) — avoid complex multi-action requests, since AI video tools handle one clear motion idea per clip best
- Avoids real people, copyrighted characters/brands, and any on-screen text
Output as a numbered list, formatted as:
[Timestamp/script section it corresponds to] — [Full video prompt ready to paste into Meta AI] — [Suggested clip length]
Also flag any moments where a subtle looping ambient motion clip (like slow-moving clouds, rippling water, gentle particle drift) could work as a flexible background layer independent of the timeline.
July 10, 2026 402 4
