Image to 3D Workflow
Use an AI 3D model generator workflow on the Astorie canvas to turn a clear 2D reference image into a rotatable 3D preview. Start from an object or character still, connect an Image node to a 3D node and inspect proportions, hidden surfaces and material cues before deciding how the asset fits your scene.
An image-to-3D step infers depth and unseen geometry from a single 2D picture. Comparing an unprepared source image with a prepared canvas reference shows why input clarity determines how much of the 3D preview you can trust.
Use an existing object image or generate a clean 2D reference on the canvas before adding the 3D stage.
Place or generate a well-lit image of a single object or character in an Image node. Avoid crowded backgrounds, heavy occlusion and cut-off edges so the silhouette is unambiguous.
Add a 3D node on the canvas and connect the Image node output to its input port. The documented 3D node workflow requires an image input rather than a direct text connection.
Choose an available image-to-3D model and quality tier in the 3D node, then run generation. Wait for the rotatable preview to finish processing on the canvas.
Orbit around the generated preview to check front, side and rear angles. Capture useful viewpoints for downstream Image or Video nodes, or verify the available output controls in your workspace.
A generated 3D preview is a spatial draft for blocking and visual reference. Inspect three areas before using captured angles in downstream shots.
Rotate the preview through a full turn and compare its width, height and depth with the source image. Look for flattened limbs, stretched depth or extra geometry attached to the base.
Any side that was not visible in the 2D image is an inference. Check the back, underside and recessed areas closely before planning a camera move that exposes those angles.
Strong highlights or cast shadows in the input photo can bake into the surface appearance. Inspect how the model reads from multiple angles when your scene requires a different lighting direction.
Use this summary to align your source image, canvas steps and project handoff before generating.
| Input | One clear 2D object or character image connected from an Image node to a 3D node. | |
|---|---|---|
| Output | A rotatable 3D preview on the canvas for spatial inspection and angle capture; verify any file export options directly in your active model node. | |
| Review | Check front-to-back proportions, inferred rear surfaces, silhouette edges and baked lighting against the original image. | |
| Plans and credits | Generation uses workspace credits according to the selected model and tier; review current plan access on the pricing page. | |
| Rights and exports | Use reference images you have permission to process and confirm downstream delivery requirements before relying on a generated preview in production. |
Combine 3D previews with documented node workflows and model references on Astorie.
Review the required Image-to-3D node connection, setup steps and quality tier notes.
See how single-subject 3D previews support camera blocking and downstream reference frames.
Check current plan tiers, trial access and credit rules before running multi-step 3D tests.
The documented 3D node workflow converts a 2D visual reference into a 3D representation. If you are starting from a written concept, generate and approve a 2D still in an Image node first, then pass that image into the 3D node.
Use a sharp, evenly lit image of one subject with a clean silhouette and minimal background clutter. Front or three-quarter views that show the main volume clearly leave fewer hidden areas for the model to guess.
Public Astorie pages describe 3D node outputs differently across docs and model guides, so treat the result first as an on-canvas 3D preview and spatial reference. Check the actual controls on your selected model node before planning a workflow that depends on exported mesh files, clean topology or UVs.
An image-to-3D model focuses on one object, prop or character that you can orbit from the outside. When you need to step inside an environment and explore viewpoints across a room or exterior location, use a scene-level world workflow instead.
Place a clean subject image on the canvas, connect a 3D node and inspect the rotatable preview from every angle your project needs.
