Skip to content

Bringing your work in

Uploading files

Drag anything supported onto the canvas and it lands where you dropped it, uploads in the background, and joins the library.

Drag a file from your desktop onto the canvas. It lands where you drop it, uploads in the background, and is ready to feed into a tool as soon as the thumbnail resolves.

There is no import dialog and no upload queue to watch.

Kind Formats
Images .png .jpg .jpeg .webp .svg
Video .mp4 .mov .webm
3D .glb .gltf .obj .fbx .stl

Audio is not on that list. Sound arrives on a board by being generated there, from the sound tools, rather than by being dragged in.

  • Double-click empty canvas and pick Upload from the ring for a file picker, or Assets to pull something you already have
  • Paste an image straight from the clipboard
  • Paste a URL to get a link card, a video player, or a live web page

Two places at once.

The file appears on the board where you dropped it, and it joins the library for the whole workspace. That means a reference uploaded on one board is available on every other board in that workspace without uploading it again.

It is also read on arrival. A vision model writes a description, tags, dominant colours and a style reading, which is what makes your own reference material findable alongside the things you generated. See automatic metadata.

The storage check runs before the transfer starts, so a full workspace tells you immediately rather than failing at the end of a large file. Nothing already stored is at risk; the workspace simply cannot take more until room is made.