Generating is best for exploration. Uploading is best when a specific person, product, frame, or recording must remain exact.
Start from scratch when you need ideas, styles, atmospheres, or options. Upload your own material when precision matters.
Most real projects combine both: uploaded product shots or portraits plus generated backgrounds, scene extensions, voice, music, or edits.
What you need
- A decision about what must be exact and what can be generated freely.
Step by step
- 1
List fixed elements
Write down what cannot change: face, product, logo, location, voice, or footage.
- 2
Generate flexible parts
Use AI for backgrounds, style tests, scene variants, and supporting visuals.
- 3
Upload exact anchors
Bring in the source materials that must remain recognizable.
- 4
Use anchors as references
Connect uploaded materials to new generations as references or start frames.
- 5
Save useful outputs
Keep strong generated variations in the library for future projects.
Simple decision rule
- Generate
- Use when you need visual direction, variants, atmosphere, or a fresh asset.
- Upload
- Use when identity, product shape, brand details, or existing footage matters.
- Combine
- Use uploaded anchors and generated variations inside the same project.
FAQ
Is uploading always more accurate?
It is more accurate for identity and source material, but generated assets may be better for exploration.
Can one project use both?
Yes. That is often the strongest workflow.
