A few weeks ago at the Beijing International Film Festival, an AI-generated short film called Peony Chronicle won the AIGC Short Film Unit. It tells the story of the late-Tang poet Yu Xuanji, rendered entirely in the texture of rice paper—her translucent robes, her falling hair, every fold of fabric floating in fine paper fibers. The film looks nothing like typical AI output. The kicker: it was made by three people, with a budget of 2,000 RMB (about $280), over two months.
The tool behind it: LibTV’s new Team Edition. I had already written about LibTV before, but this release marks a fundamental shift—from individual tinkering to collaborative production. To understand how it works, a friend and I decided to build a 30-second Pixar-style animation of a scene from The Hitchhiker’s Guide to the Galaxy: Arthur Dent’s last cup of tea before Earth is demolished by the Vogons.
The first thing you notice in the Team Edition is the shared canvas. It still looks like the personal version—an infinite dark workspace—but now the top-right corner shows a sync status, a list of online members, and live avatars. You are no longer alone. My friend handled the Vogon captain; I did Arthur and Ford Prefect. Each of us generated a standard-angle character image, right-clicked to “create team subject,” and let LibTV auto-generate nine additional angles. The whole process happened on the same canvas. I could see his cursor, which node he was editing, and a “editing” badge appeared on that node. We never overwrote each other’s work—just like editing a shared document on Feishu.
The real power is the team subject library. As soon as my friend finished the Vogon captain, it appeared in my library under “characters.” Any team member can drag any subject into their own scene—ensuring that the same character appears consistently across all shots. We also created a shared image preset for the Pixar look, so every frame kept the same visual tone. This ability to maintain character and style consistency across a distributed team is what separates a professional AI pipeline from garage experiments.
Once the subjects and style were locked, the rest became almost automatic. We fed a short plot description into the script node (GVLM 3.1), and within seconds it generated a full storyboard table: each shot included scene description, camera movement, characters, expression, action, lighting, color palette, sound effect, shot type (medium, low angle, close-up, etc.), and resolution. The detail rivals a traditional director’s storyboard, but it took 30 seconds instead of three days.
What does this mean for the industry? Hollywood is still battling over whether to allow AI in filmmaking; meanwhile, Chinese studios and independent creators are already producing festival-quality shorts with a fraction of the budget. Traditional 3D animation can cost $10,000 to $50,000 per minute. With tools like LibTV, teams of three to five people can achieve similar results for under $500 per minute. Compared to Runway or Pika, which focus on solo prompt-to-video, LibTV’s collaboration layer is unique—it mimics the iterative workflow of real production, where directors, artists, and editors need to iterate in tandem.
For independent creators, this is a watershed moment. Imagine a micro-studio of three friends in different cities, each responsible for one character, all working on the same scene in real time, with a shared library of assets that never drift. The era of the lone AI video artist is ending; the studio is going remote. The next big blockbuster might not come from a giant studio—it might come from three people, a laptop, and a $300 budget.
If you are a filmmaker, animator, or storyteller, now is the time to experiment. The tools are cheap, the collaboration is real, and the audience—at festivals like Beijing’s AIGC unit—is already paying attention.