| name | scrollclaw-animate |
| description | Animate first frames into A-roll talking head clips. Uses Sora 2 via fal.ai (primary) with Kling 3 auto-fallback when Sora is unavailable or sunset. |
| metadata | {"openclaw":{"emoji":"🎥","user-invocable":true,"triggers":["animate clip","sora animate","generate clip","a-roll","talking head clip","sora video","kling animate","kling a-roll"]}} |
Animate (A-Roll)
Turns first frames into talking head clips with synced lip movement and audio. Image-to-video is the default — text-to-video is the fallback.
Fallback chain: Sora 2 (fal.ai) -> Kling 3 (fal.ai) -> Kling 3 (Replicate). Use --provider kling to skip Sora entirely (for when Sora is sunset).
Prerequisites
- First frame approved (from
/first-frame)
- Script with
[A-ROLL] segments tagged (from /persona)
Motion Prompting
Read references/motion-prompting.md for the structured prompt format. Use labeled fields — not prose paragraphs:
- Camera — handheld energy, micro-shake, slight reframe
- Subject — keep generic (first frame defines appearance). Detailed facial descriptions trigger content safety filters
- Dialogue — include actual script lines. Sora generates synced lip movement + audio
- Audio — describe the RESULT not the gear. "Clean natural podcast audio, voice close and present, subtle room tone" works. Naming specific mics doesn't.
- Environment & light — practical light, deep focus, real-world setting
