Workflow
-
Define the job:
- audience
- surface
- output format
- desired viewer action
- brand role
-
Decide the source of truth:
- product UI: code and browser capture
- static visual direction: imagegen/GPT Image
- cinematic generated clip: Sora, Grok, Veo, Luma, Runway, or Higgsfield
- exact typography/layout/video package: Remotion
- export/compression: ffmpeg
-
Write the still-frame prompt before the video prompt.
-
Generate or select style frames.
-
Write the video prompt:
- source frame
- camera
- subject motion
- environmental motion
- audio
- beginning/middle/end
- constraints
-
Route providers:
- Sora for OpenAI-first cinematic video and image-to-video.
- Veo for high-fidelity controlled video, native audio, first/last-frame workflows.
- Grok Imagine for fast social/audio-synced clips and stylized tests.
- Luma for cinematic exploration and creative-agent workflows.
- Runway for production editing and existing-video transformation.
- Higgsfield for creator/social/ad variants.
- Remotion for final deterministic finishing.
-
Finish:
- captions
- title cards
- logo lockups
- CTA
- aspect-ratio variants
- GIF/MP4/WebM exports
-
QA:
- first frame works as still
- crop works on target channels
- text is readable
- loop resolves cleanly
- audio is intentional
- file size is acceptable
- provenance is complete
Image Prompt Shape
Use case:
Asset type:
Primary request:
Input images:
Scene/backdrop:
Subject:
Style/medium:
Composition/framing:
Lighting/mood:
Color palette:
Materials/textures:
Text (verbatim):
Constraints:
Avoid:
Video Prompt Shape
Output:
Aspect ratio:
Duration:
Provider candidates:
Source frame:
Subject:
Setting:
Camera:
Subject motion:
Environmental motion:
Audio:
Beginning:
Middle:
End:
Style:
Constraints:
Avoid:
Output Shape
Return:
- creative diagnosis
- recommended workflow
- provider routing
- still-frame prompt
- video prompt
- finishing plan
- QA checklist
- provenance template