Skip to main content

Sets

A Set is the environment or location where your scenes and shots take place. Sets can be AI-generated from text descriptions, imported from 3D tools like Blender, or built from scratch in the editor. They provide the backdrop for digital actors, props, and camera action, and ensure visual continuity across your project.

Types of Sets​

AI-Generated Sets — Describe a location in plain text and ACT3 AI generates a matching environment. Examples: "Ancient temple interior at sunset," "Modern corporate office with city view," "Abandoned spaceship corridor, emergency lighting."

Template Sets — Pre-built environments ready to use immediately: office, street, forest, sci-fi lab, restaurant, warehouse, and dozens more. Templates are a fast starting point you can customize.

Imported Sets — Import 3D models and environments from Blender (.blend, FBX, OBJ, GLTF) or Unreal Engine (USD) for advanced worldbuilding and full 3D control.

Custom Sets — Build from scratch in the Top-Down View by placing props, furniture, and background elements on a virtual stage.

Creating a Set​

  1. Open the Editor and go to Set Management from the sidebar
  2. Click Create New Set
  3. Choose a method: Generate from description, choose a template, or import a file
  4. If generating: enter a description, select lighting and time-of-day settings
  5. Customize props, furniture, and environmental details
  6. Place digital actors, cameras, and lighting rigs in the space
  7. Save the set to your Asset Library for reuse

2D and 3D Views​

Sets in ACT3 AI work in two modes:

2D Layout View — A bird's-eye planning surface. Drag actors, cameras, and props from the media library onto the stage. Use this for fast blocking and camera positioning. 2D layouts use no credits.

3D Environment View — A real-time perspective view of the set with lighting, textures, and depth. Switch to 3D to check angles, depth of field, and how lighting falls on actors. 3D renders and exports consume credits based on complexity.

You can switch between 2D and 3D at any time. Use 2D for initial planning and 3D to validate your decisions before generating video.

Top-Down View for Set Blocking​

The Top-Down View is specifically designed for camera and actor blocking:

  • Place camera tokens at positions with orientation arrows
  • Draw movement paths for dolly and tracking shots
  • Position digital actors with directional markers
  • Trace character movement paths across the set
  • View multiple camera setups simultaneously to check coverage

This prevents continuity errors and saves credits by planning before rendering.

Set Continuity​

For episodic content or projects with recurring locations, save your sets to the Asset Library and reuse them across multiple scenes or projects — see create a set for the full workflow. A saved set retains:

  • Prop placement and arrangement
  • Lighting setup
  • Camera preset positions
  • Actor starting positions

Apply the same set to multiple scenes to ensure visual consistency without rebuilding from scratch each time.

Integrating Sets with the Production Pipeline​

Sets connect to the rest of your workflow:

  • AI Wizard — Suggests set types based on your story beats and scene descriptions
  • Script Editor — Scene headings are linked to set assignments
  • Rendering System — Generates the final environment visuals in high resolution

Import Formats​

FormatUse
.blendBlender projects with full material support
FBX3D characters, props, and environments
OBJLightweight 3D objects
USD / USDZComplex scene hierarchies and AR-ready assets
GLTFEfficient web-ready 3D scenes

Procedural City Sets​

The City on Rails set type generates a full city block procedurally. In the Set editor, choose Set Type → City on Rails. Define building types for each location (restaurant, store, park, office) and the overall style (modern, historical, dystopian, futuristic). ACT3 AI assembles a walkable urban environment you can use across multiple shots with consistent street-level geography.

Top-Down Canvas for Blocking​

The Top-Down Canvas provides a bird's-eye planning surface — similar to Figma — for placing characters and cameras before generating any video.

  1. Click the Top-Down tab in the editor
  2. Pan and zoom freely across the set layout
  3. Drag character icons onto the set layout and draw movement paths using the spline tool
  4. Place virtual cameras and see their field-of-view cones
  5. Mark camera positions with orientation arrows to define shot direction

This blocking plan feeds directly into the shot data for generation. Planning here saves credits by reducing trial renders.

3D Ground Alignment for 2D Backgrounds​

When using a 2D image as a set background, use the 3D Ground Alignment tools to define the ground plane and vanishing points. This allows ACT3 AI to correctly position 3D characters in perspective against the 2D background, giving the illusion of depth and proper scale matching between the character and the environment.

Browsing the Set Library​

Open the Sets panel and use the search bar to find sets by keyword. Filter by:

  • Set type — 2D image, 3D Blender, Procedural
  • Location category — organized by Country → City → Campus or Environment → Interior → Exterior
  • Usage — sets used in the current project

Sort by title, date added, or community rating. Mark sets as favorites for quick access to your preferred environments.

Reusing Sets Across Scenes​

Sets are saved to your organization's Set Library and can be assigned to any scene across any project. Assign the same set to multiple scenes to ensure visual consistency in that location throughout your film — the same hallway in Act 1 will look identical when it reappears in Act 3.

Best Practices​

  • Use consistent sets across related scenes for visual continuity
  • Save approved sets to the Asset Library before making changes
  • Plan camera paths in Top-Down View before committing to full renders
  • Apply a global lighting preset to unify the look across different sets in the same project
  • Use 2D blocking for all planning, 3D only for final visual confirmation

Troubleshooting​

Set not loading correctly — Check that all asset files are included when importing, and that formats are supported. For Blender files, save with linked textures packed.

Actors not visible in set — Verify actor placement in Top-Down View and check that the scene assignment is correct.

Render too complex, running slow — Simplify set elements for preview renders. Use proxy assets at lower resolution during the creative phase.

Inconsistent visuals between shots — Apply a global LUT or color grading profile to ensure all shots in the same set match.

Spots and Rooms​

Sets support two layers of sub-organization:

  • Set Spots — named positions within a set (e.g., "front door," "left window," "center stage") that actors and cameras can reference by name in shot descriptions
  • Set Rooms — bounded sub-areas within a larger set (e.g., "Kitchen," "Living Room," "Corridor") with their own descriptions, lighting, and spot definitions

Use compass directions (North, South, East, West based on the Top-Down View canvas) or frame directions (frame left, frame right, foreground, background) to describe positions within a set or room.