Lip Sync
Voice-driven performance
Feed a take and an audio track; the room re-times mouth shapes to the performance. The identity lock carries the face, the model handles the syllables.
In progressRoadmap
Built in public, on the bench.
Moviola grows room by room. Here is what already ships, what is being bolted together right now, and where the architecture is headed — no vaporware, dates only when they're real.
On the bench now
Three rooms under construction
VFX
Explosions and transformations
Effect presets — fire, smoke, dissolve, morph — translated per model the same way camera presets are today. Stack them on any finished take.
In progressWorkflow
Chained multi-step pipelines
Wire rooms together: storyboard → identity lock → render → upscale → lip sync, as one saved pipeline you can rerun with a new prompt.
In progressThe method
Every room is a plugin. The bench is forever.
plugin architectureno core rewritesrooms keep arrivingplugin architectureno core rewritesrooms keep arrivingplugin architectureno core rewritesrooms keep arriving
Already rolling
The shipped ledger.
Four rooms live today, all free, all in the same Docker image you can pull right now.
Image Studio
Text→image, reference-driven edits, upscale & matte chaining.
Video Studio
Text or start-frame motion, 8 camera presets, stack three per take.
Storyboard
Shot-by-shot planning with per-shot identity and batch render.
Identities
Reusable characters that lock across studios and models.
Watch the rooms get built.
The whole thing happens in the open — commits, experiments and the occasional revert. Star the repo or check the changelog every few reels.