I’m looking for someone to help build a structured AI video generation workflow for TikTok Shop style videos.
The type of videos I want to replicate are these examples from my Drive:
https://drive.google.com/drive/folders/1gRkxwPBqy0QKEv3IFqE2wPZTGlaUHjcf?usp=sharing
I’m planning to use Veo 3 inside Flow for the generations (but I’m open to discussing better tools/workflows if you have experience with alternatives).
As you can probably notice, the videos follow a very specific and consistent pattern:
• static camera
• natural body motion
• smooth continuous movement
• realistic TikTok aesthetic
• coherent motion between clips
Since Veo currently generates a maximum of around 8 seconds per clip, and the reference videos are usually around 20 seconds long, the workflow would need to generate multiple connected clips.
The prompts need to be interconnected so the motion, body positioning, lighting, framing, and overall aesthetic remain coherent across all generations.
Most likely, the last frame of each generated clip would also be used as the starting frame for the next generation to maintain continuity.
What I need is:
• a reusable prompt structure/system
• around 3 interconnected prompts for sequential video generation
• prompts designed specifically for realistic TikTok try-on/body movement videos
• an easy way to swap or modify movement patterns while keeping consistency
Ideally, I’m looking for someone experienced with:
• Veo / Flow
• AI video generation
• prompt engineering
• motion consistency
• realistic TikTok-style outputs
• generative video workflows
If you already have experience building structured generation pipelines like this, please send examples or explain your approach.
Apply tot his job
Apply To this Job