AI Tiny DeskGuide
Tiny Desk AI prompt: an example for video editing
Updated November 1, 2026
A video-edit workflow can combine reference footage with portraits. The example below illustrates a prompt structure; it has not been independently tested with every model or provider.
Using reference material
A reference video can give an editing model a source for camera movement and timing. Check the provider’s input requirements and whether you have permission to use the footage and portraits.
Illustrative prompt
Replace selected performers in the reference video with the people in the supplied portraits. Keep the camera movement and timing where possible. Review the generated result before sharing.Before using an editing model
- Model: check the provider’s documentation for supported video-edit and reference inputs.
- Format: confirm your chosen model supports the desired aspect ratio and resolution.
- Video length: check the provider’s limits for the input and requested output.
- Audio: check your rights and whether the provider preserves sound in the exported file.
Manual reference workflow vs. pre-configured stage template
| Workflow step | Manual video-edit workflow | AI Tiny Desk |
|---|---|---|
| Reference clip & shot cuts | Supply your own reference clip and align shot boundaries manually | Pre-configured 4-shot stage template with cut re-alignment |
| Unchanged-face check | Inspect output manually; failed runs consume model credits | Automated face check; failed jobs automatically return credits |
| Audio & social mute fallback | Sync audio track separately in an editing app | Provides both audio-synced MP4 and silent MP4 backup |
| Setup effort | Write role prompts, map reference images, and stitch exports | Select 1–5 portraits in the browser and download when ready |
Tiny Desk AI guide: make your video in 4 steps — Photo rules that actually matter, how the 5 band spots work, and how to fix a weak render.
Tiny Desk AI prompt FAQ
What is the best prompt for the Tiny Desk AI trend?
A prompt can map each reference portrait to a role and describe the desired motion. Reference syntax and results vary by provider.
Can ChatGPT or Sora make the Outstanding Tiny Desk video from text?
Text-only output may differ from a specific reference performance. Check the selected provider’s current capabilities and content rules.
Why does the lead singer's face glitch in the first second when I do it myself?
Results can vary with camera angle, portrait and model. Check the provider’s guidance and review the output before sharing.