AI Video Studio
One studio for Sora, Veo and LTX, with an asset library that compounds.
The question I was chasing
Why should three frontier video models mean three tabs and three ways of thinking?
Sora, Veo and LTX do not agree on much: different strengths, different inputs, different ways of failing. The question was not whether I could call a video API. It was whether one coherent place to work across all three could exist.
Working across the providers by hand means re-uploading references, losing track of what was generated where, and starting from scratch each time. I wanted a studio that held the whole workflow in one place.
The constraints
Three video services, each with its own controls and quirks, had to sit in one interface that keeps each model's full range rather than flattening them. Generated assets had to be reusable across projects rather than trapped in one render. And the whole thing had to stay coherent as models and controls shift underneath it. This is a moving target by design.
The decisions that mattered
An asset library is the connective tissue. Still images live there and can be pulled in as references across any project, so work builds on itself instead of resetting.
I exposed the full set of controls in a custom interface rather than hiding behind presets, because the point of a studio is control. The cost is that every control I surface is one more thing to keep current as the services move underneath it.
I treat the three video models as interchangeable engines for a render rather than three separate apps bolted together. The workflow is the product; the model is a choice inside it.
Where it landed
Three models render from one place, picked per shot. Generated images sit in the asset library and get pulled into any project as reference material. The full set of controls on all three models is available in the interface rather than hidden behind presets.
Part of the Rolling Waves work archive.