Hey everyone! I’m Moheb, an AI filmmaker from Egypt. I build controlled, repeatable pipelines (Nano Banana → Kling → DaVinci Resolve) focused on shot continuity and character consistency — not random generation. Excited to share breakdowns and learn from this community. What’s everyone’s biggest continuity challenge right now?
For more info, visit My Website
Welcome, Moheb. The biggest continuity challenge for me is no longer only character identity. It is maintaining the exact spatial and temporal state of a scene across different camera positions.
A face can remain recognizable while the room geometry changes, an object moves to the wrong side, a doorway appears, or the model silently jumps from a pre-action state to a later moment.
What has helped me is separating a scene into approved states—before the action, during the action and after it—and creating a stable anchor frame for each important state. New shots are then derived from the nearest approved anchor rather than repeatedly generated from the original description.
I also keep a fixed scene block for location, character positions, lighting and objects, while changing only the camera section of the prompt. When a result drifts, I return to the last stable anchor instead of trying to repair the drift through another generation.
I’m curious how you handle room geometry when moving to a substantially different camera axis.