AI Film Hackathon in Tokyo: Building an AI Short Film with Magnific in Two Hours
I attended the AI Film Hackathon in Tokyo at the Fujitsu Kawasaki Tower and used the AI animation tool Magnific for the first time. It ships with MiniMax H3 series models and Google Flash 3.8 TTS, while an AI Agent automatically operates the canvas to build the Base layer. At a roughly five-hour event I finished seventeen or eighteen audio-driven clips totalling nearly two minutes in under two hours, and the promo CM I made for a pet-care App won an award. I also sense coding agents giving way to more expressive AI products.
AI Film Hackathon in Tokyo: Building an AI Short Film with Magnific in Two Hours
Today I attended the AI Film Hackathon in Tokyo. The venue was the Fujitsu Kawasaki Tower.
Thanks to the Organizers and to Magnific
First of all, I want to thank the organizers very much. They gave me the chance to introduce Magnific, an outstanding AI animation editing tool, to everyone. It is a tool similar to TapNow. Today was my first time using it, and it was extremely easy to pick up — I was up and running in about ten minutes. Although it runs in the browser, the controls are remarkably smooth.
A Rich Model Lineup and Very Cheap Audio Generation
Magnific offers a particularly large number of models. Above all, the MiniMax series I have been following recently — MiniMax H3, MiniMax H3 Max, MiniMax H3 Max Turbo and so on — includes every model I really wanted to try.
It also has a sound generation mode powered by Google's latest Flash 3.8 TTS. Generation is very cheap, costing only about four or five credits each time, which is excellent value.
An AI Agent That Operates the Canvas Automatically, Saving Huge Amounts of Time
Another especially powerful feature is the AI Agent that automatically operates the canvas. It is genuinely impressive: once I hand it a structured story, it receives and understands it on its own, then builds out the entire Base layer I need on the canvas, saving a great deal of time. After that I just check it myself and start producing each piece one by one.
Efficiency: A Five-Hour Event, Finished in Under Two Hours
The event lasted roughly five hours, but I finished all of my work in under two hours. The videos I generated add up to nearly two minutes. Its ability to handle longer pieces is such that this breaks down into seventeen or eighteen short clips, and every one of them is audio-driven. The tool's results really are excellent.
The organizers were also extremely generous: they gave every participant 20,000 credits, for which I am truly grateful. Even after finishing a video of just over two minutes made of seventeen or eighteen clips, audio included, I still had credits left over. So this tool is genuinely a good one.
The Event's Two Themes and the Product CM I Made
Today there were two themes: one was to make a short film, and the other was to make a product CM. It happened that in August and September, together with my teammates, I joined another hackathon and built an App for taking care of pets. That is a small product. Taking advantage of this event, I made a promotional CM for it. Completely unexpectedly, my piece actually won an award and was selected. I was delighted — even more delighted — and I am truly grateful to the organizers.
A Few Closing Thoughts
Finally, the VibeCoding was already quite lively in September and October of last year. But since this autumn, including at these events in Tokyo, I can clearly feel that animation generation and AI generation are taking off. So I believe that between this year and next, coding agents will gradually settle down, while animation agents, music agents, and other more expressive AI products will rise to the fore. I also look forward to the prosperity of Japan's IT technology market as a whole.