Follow a time traveler as she keeps getting sent to the wrong era when a new time dispatcher has his first day on the job. π¬ Tutorial: β’ Make an Entire AI Movie With Just One AI T... I've been re-making this AI comedy every year or so to track how well generative AI is improving and how storytelling and filmmaking is changing along with it.
Workflow & Recipe
The project was built around a self-imposed constraint to use a single all-in-one platform, OpenArt AI, to test how far integrated generative tools have progressed. The concept was defined first as a recurring time-travel gag centered on a dispatcher error, allowing the story to tolerate visual inconsistency and lean into comedic timing rather than realism. All images, video clips, and dialogue were generated directly inside OpenArt using a mix of internal models such as Flux, Nano Banana, Veo, WAN, and Kling, chosen scene by scene based on what each model handled best. Dialogue and performances were generated within the same environment, accepting limitations in voice acting as part of the experiment, while music was created separately using Suno. The final film was assembled by selecting the strongest generations, pacing them for comedic rhythm, and using repetition and escalation to emphasize the evolving quality of AI outputs over time, making the film both a narrative piece and a longitudinal experiment in generative filmmaking.