Skip to content

3D

3D on a 2D canvas

Real geometry sitting beside your images means you can decide the angle before you generate, instead of describing it and hoping.

You have a product shot to make and the client wants it from a specific three-quarter angle, slightly above, with a long lens. Describing that to an image model is a sentence you keep rewriting.

With the object on the canvas as real geometry, you set the camera and the model gets the answer instead of the question.

Exact framing. Orbit the object, pick the focal length, light it. The view you settle on is a real render of a real object, and every proportion in it is correct because it came off the geometry.

Consistency across a set. Six shots of the same object from six angles are six views of one mesh rather than six independent guesses that will not quite match.

A starting point that is already right. Feed that render into an image tool and the model is enhancing something accurate rather than inventing something plausible.

The board is an isometric 3D space, and most of the time that is invisible. Images, notes and cards behave exactly as they would on a flat canvas.

The difference shows when a mesh lands. It sits in the same space as everything else, so:

  • You orbit it without opening a viewer
  • A comment pins to a point on its surface and stays there as it turns
  • It scales, groups, frames and locks like any other object
  • A render of it lands beside it rather than in a downloads folder

Honest limits, because a shared 3D canvas is not free:

  • Heavy meshes make the board heavier for everyone, since the board is shared. Remesh before you import when a scan is millions of triangles
  • The board camera does not orbit. The isometric angle is fixed so that everyone’s view of the board matches. Individual objects orbit; the board does not
  • A thin laptop will feel it on a board with a lot of geometry, because the rendering is happening on your machine