Video editing has always meant working around the footage you captured.

AI is changing how much of that footage can be reimagined without rebuilding the entire edit.

ToneFrame explores a more direct workflow: upload a shot, describe what should change, and generate a new version while preserving the original motion and performance.

Relight a scene. Remove distractions. Restyle the wardrobe. Transform the environment. Create a different atmosphere. The direction begins with a prompt rather than a timeline full of masks and layered controls.

I’m currently testing Google Gemini Omni Flash and Kling 3.0 to compare visual consistency, motion preservation, resolution, generation speed, and cost. The goal is to determine which model delivers the strongest foundation for production-quality, prompt-led video editing.

ToneFrame is still in development, but the idea is simple: keep the shot, change its world.

See the project at toneframe.com