Storyboard planner
Write a brief, get a storyboard, and watch the film. One shot at a time.
write
Advanced
project
shots
Read and edit each shot. Changes go into the film automatically.
validation errors
storyboard JSON / YAML (operator)
storyboard JSON (validated)
storyboard.yaml (bundle-ready)
raw model output
Preflight (operator)
preflight
validator + scene shape + cast readiness. errors used to gate Assemble; they still list here.
Audio (optional score or narration)
audio bed + beat timing
optional. generate a track or upload your own. set BPM, then snap shot durations to musical phrases so the cuts land on the beat. Skip this and the film stays silent.
plan or load a storyboard first; audio + beat timing attach to the storyboard's scenes. you can upload or generate a track once a plan exists.
generate music
describe the track: genre, mood, instrumentation, era. duration is up to the model; expect ~30-60 seconds. "suggest from video" drafts a prompt matched to your storyboard's mood, tempo, and energy arc (auto-filled the first time you open this step).
generate narration
spoken narration for the audio bed. leave blank to let the module derive a script from your storyboard, or write the lines to synthesize.
upload your own track
mp3, wav, aac, m4a, ogg up to 32 MB.
Who is in this movie
This film can have up to four people. Pick from the ones you already have, or type a new name.
In this film
No saved characters yet. Type a name and hit Add.
Assemble
Confirm who is in the film, then assemble. A portrait per person is enough for a first film.
per-scene start keyframes (optional)
attach an authored keyframe to drive a scene's GPU (Wan) motion from that exact frame; scenes left empty render normally.
Render this film
Stills or motion, which look (the backend), how good, then Render this film.
Finish this film
Motion already talks when the model can. Each shot animates toward the next still. MuseTalk is only if you want to replace the mouth.
Voice is chosen once on Cast for each person. We fill this lock from that so every shot is the same speaker. Edit only if you need a tighter description.
title + credit cards (optional)
Text for the opening title and end-credit cards, used when a title-card film.finish module is installed. Leave blank for no card.
Advanced
module settings (optional)
Controls are rendered from installed modules' config_schema. Empty fields use each module's default. Quality tier above is injected into keyframe and motion modules automatically.
expert: raw JSON
optional JSON in the module wire shape:
{ "motion_backend": "seedance", "config": { "cloud-keyframe": {…}, "seedance": {…} } }.
wins over the structured fields above on conflict.