How it works
- Define your elements — Create elements on the canvas for each character, prop, or scene in your project. Assign each a name, type, and reference image.
- Describe your shots — Tell the AI agent what happens in each shot, referencing elements by name.
- Generate with consistency — The AI uses your element references to maintain visual consistency across shots.
Setting up elements
Each element on the canvas has three key properties:- Name — A label the AI uses to identify this element (e.g., “Detective”, “Red car”, “Office”)
- Type — Character, prop, or scene
- Reference image — The visual reference that defines how this element looks
Create an element
Upload an image to the canvas or generate one. Convert it to an element by setting its name and type.
Define its role
Choose the element type:
- Character — People, creatures, or animated entities that appear across shots
- Prop — Objects that need to look consistent (vehicles, tools, furniture)
- Scene — Backgrounds or environments
Generating multi-shot sequences
Once your elements are set up, describe your sequence to the AI agent: “Create a 3-shot sequence: Shot 1 — Detective walks into the Office. Shot 2 — Detective examines the Red car. Shot 3 — Close-up of Detective’s face looking surprised.” The AI generates each shot while keeping the Detective, Office, and Red car visually consistent.Aspect ratios
You can specify the aspect ratio for your generated videos. Common options include 16:9 (widescreen), 9:16 (vertical/mobile), and 1:1 (square).Audio
Multi-shot generation supports audio generation. If your character elements have voice profiles assigned, the AI can generate videos with matching voiceover or dialogue.Credit costs
Multi-shot video is priced per second of generated content:| Duration | Credits |
|---|---|
| Per second | 280 |
| 3 seconds (minimum) | 840 |
| 5 seconds | 1,400 |
| 10 seconds | 2,800 |