One story → scripted scenes → consistent 3D characters & voices → stitched film
How it works
Paste a story idea — the AI writes a director-grade script: max 3 dialogue lines per scene, a full ACTION field per scene, Scene 1 built as a scroll-stopper.
Pick the number of scenes — each scene is one 15s-max H3-Max generation. The total film runtime shows in minutes as you choose.
One locked cast: the same reference images + cloned voices are bound to every scene, so characters stay consistent across the whole film.
Scenes run as a queue (pause / retry any failure), then Stitch All joins them in order into one film.
1 · Story & Script
2 · Cast — Faces & Voices (locked for all scenes)
Character Reference Images (cast order: Image 1, Image 2, …)
Cloned Voices (cast order: Audio 1, Audio 2, …)
Reference Audio Files (optional extras)
Character ↔ Asset Legend (filled by the script)
Write the script first — characters get @ImageN / @AudioN tags in cast order.
3 · Film Length
—
Choose the scene count to see the total film length.