{{ p.context }}
{{ cueList(p) }}
{{ modeInfo.hint }}
Change one thing per pass and say what is now true of the scene, not what to do to the footage. “It is evening now” works; filter language like “apply a 35mm grade” has been reported to do nothing. Everything you don't protect is free to drift, so list what must stay.
Describe the files you'll upload at generation time — you don't need them here. Each one gets a number in upload order and a declared role. Keep this list in the same order you upload. Video and audio are each capped at 15 s total.
No references yet. Add the first file you plan to upload.
One frame fixes the opening or the ending. Two frames make H3 fill in the motion, light and sound between them. It won't invent cuts in this mode, so write the bridge as a single continuous move.
Anything that has to stay the same all the way through: a person, a product, a location. Name the two or three details that make it recognisable rather than writing "the same woman".
Nothing defined yet. Add whoever or whatever carries across the clip.
Four things the manual asks for: subject, place, event, style. Write what the camera sees, not what it means.
Describe the clip in playback order. One main camera idea per shot. Dialogue goes in word for word — a paraphrase produces mush. The strip below stays pinned while you scroll: click a block on the timeline or a number to jump to that shot. Use the bar at the bottom to drop a subject or file token wherever the cursor sits.
{{ s.shots.length - matchCount }} shot{{ s.shots.length - matchCount === 1 ? '' : 's' }} hidden by “{{ findQuery }}”. Show all
Give the performance a cause, then describe brief, observable facial changes. A named emotion alone is weaker than context plus eyes, brows, mouth, breath, and a clear before/during/after progression.
Each item has its own phase, exact anchor, duration, side, and intensity. Use the arrows to control the order within a phase.
Prompt: {{ microCuePreview(sh,cue) }}
No individual expressions yet. Choose a recipe or add the first movement manually.
The thread demonstrates bracketed cues during speech. Put a cue at the beginning or after the sentence, or anchor it before exact words in the dialogue.
No shots yet. Even a single continuous take is one shot.
Sounds tied to an action stay in the shot that makes them. This section is for the clip as a whole. Music the characters can't hear is separate again — if you don't want any, say so explicitly.
{{ formatInfo.hint }}
{{ prompt || 'Fill in the form and the prompt builds itself here.' }}
No saved sessions yet. Name this one and save it to come back to it later.
Saved in this browser only. Work in progress is kept automatically and comes back when you reopen the page — export a file to move a session to another machine.
This browser is blocking local storage, so sessions last only until you close the tab. Export a file to keep anything.
{{ lib.msg }}
Recipes set the emotional context and an ordered run of facial, breath and body cues. They cover faces and body language for adult characters; the romantic and desire recipes stay suggestive and never explicit. Pick one, then adjust every cue in the shot.
Add a shot first: recipes are applied to a shot.
{{ p.context }}
{{ cueList(p) }}
Nothing matches. Try fewer words or clear a filter, or build your own in the Create tab.
A recipe is a context plus an ordered list of cues. Build it here, or fill in a shot's acting panel and press “Save this as my recipe”.
Prompt: {{ draftCueLine(cue) }}
Plain sentences added to the shot as they are, on top of the cues.
A movement is one visible thing a face or body can do. Once added it appears in every expression menu and can be used in recipes.
Prompt: {{ mvPreview }}
Nothing yet. Use “Copy & edit” on any built-in recipe, or press Create.
None yet. Add one in Create, under New movement.
Your library is saved in this browser automatically. Export it to keep a copy or move it to another machine. Import merges into what you have and never overwrites.