A new performance starts with generation
A motion model can interpret still images and a movement reference to produce new frames. This requires checking subject identity, body anatomy, timing and interaction between the two people. A template name alone does not prove that two faces stay distinct through the whole dance.
An edit starts with existing material
A conventional editing workflow uses photos, footage, cuts and a timeline. That is useful for combining a finished dance clip with a title, changing its pacing or preparing a platform-specific export. Individual editor templates can offer additional AI features; inspect the actual template instead of assuming its name describes the full capability.
Compare the complete result
Before paying, check what goes in and what comes out. Does the workflow need one image, two images or a dance recording? Is it generating people in motion or arranging still images? Check resolution, watermark, audio, download access and what happens on failure.
A practical two-stage approach
First obtain a usable video and review both faces, hands and feet throughout it. Then finish captions, cuts and appropriately licensed music in your preferred editor. NEON DUO currently tests the upload-to-download workflow with a labelled local composition; it does not claim to outperform CapCut or provide a verified AI dance.