Glossary
The vocabulary MacroScene uses, in plain language.
- Brief
- Your short description of the story you want — a sentence or a paragraph. The starting point for everything.
- Treatment
- The brief expanded into an editable story: logline, characters, tone, and a beat-by-beat synopsis. You review and edit it before any video is generated.
- Story bible
- The consistency contract the AI keeps to across the whole film — logline, visual style, color palette, and each character's description.
- Scene
- A unit of the story set in one place, like a chapter. A scene contains one or more shots.
- Shot
- A single continuous camera take. Each shot becomes one generated clip of roughly five seconds.
- Clip
- The actual generated video for a shot — what the video model produces.
- Cut
- The jump from one shot or scene to the next. "The cut" also refers to the finished, edited film.
- Final cut
- All of a project's clips stitched together in order into one MP4 — your finished short film.
- Continuity
- How consistent characters, places, and motion stay from one clip to the next. Modes: Consistent (locked, no drift), Balanced (flow within a scene), Smooth (always flow), and Corrected (flow, then snapped back on-model).
- Character reference
- An AI-generated portrait of a character, used as an anchor so they stay on-model across every clip.
- Establishing image
- An AI-generated image of a scene's location or setting, used to keep the place consistent.
- Scene keyframe
- A composite image placing the scene's characters into its setting. Anchoring shots to it locks both the characters and the place together.
- Anchor (init image)
- The starting image handed to the video model for a shot — a character reference, a scene keyframe, or the previous clip's last frame.
- Image-to-video (i2v)
- Generating a clip that begins from an image, which is how MacroScene keeps continuity. Text-to-video (t2v) generates from a prompt alone.
- Quality tier
- Draft, Standard, or High — trades cost and speed for fidelity. Higher tiers use more capable (and more expensive) models.
- Resolution
- The output size of each clip — 480p (faster, cheaper) or 720p (sharper).
- Prompt
- The text describing a shot's action, camera, and composition that drives the generation.