Framer’s agent learns the geometry of a designed object
Framer Agents can now work with 3D transforms while keeping layers editable on the canvas.
A practical opening for dimensional identities that still behave like websites.
The Story
Framer’s latest Agent update adds support for 3D transform properties including rotation, perspective, origin, depth, and backface controls. Crucially, the generated result is not flattened: the layers remain selectable and customisable. ([framer.com](https://www.framer.com/updates?utm_source=openai))
That makes 3D less of a cinematic flourish and more of a layout primitive. It could support a product hero that turns with the cursor, a portfolio index with spatial navigation, or a brand mark that shifts across responsive states without leaving the browser.
The creative direction opportunity is to define the rules first: what rotates, what stays anchored, and what should never move. A little spatial discipline will make the effect feel authored rather than merely technical.
Studio Wensday Angle
Use 3D as a compositional system with rules, not as a novelty layer.