The visual canvas is where your work gets built. You place cards on it, one card per step, and draw connections so what one card makes feeds into the next. A set of connected cards is a flow.
Adding a card
Several ways, and where the card lands depends on which you use:
Card library, the C key, the header menu, or quick add: lands in the middle of your current view.
Right click on the canvas: lands where you right clicked.
A card's output dot, or its right click Add card: lands right beside that card.
The canvas holds still. Adding a card never re-frames your view, and new cards arrive already selected.
Undo removes one card per press.
The add menu
Adding a card opens the full preset library: style presets, image presets, and video presets. Drag the menu by its title bar to move it. The preview pins to the side at full height, so you can move onto it and read it. Picking a model or a ratio inside the menu does not close the menu.
Connecting cards
Drag from a card's output dot to the next card's input. Cards only offer the inputs the chosen model actually accepts: start frame, reference, character, control.
Connected references show as thumbnails over the card preview and update live. Past 5 they collapse into a "+N" chip.
If you change the model to one that does not accept a connected input, Visuals remaps or disconnects that connection and tells you what it did.
Selecting
Single click anywhere on a card selects it, including the thumbnail and the prompt box.
Shift or Command both work for multi-select.
Double click on the image, video, or text opens the full view.
Double click anywhere else on the card opens the edit panel.
A plain click in the prompt puts you in the prompt editor.
With several cards selected, the pill that appears has Run, Add to chat, and Download.
Aspect ratio
Everything defaults to 16:9, images and video. Connect media into a card and the card takes that media's shape, snapped to the nearest shape the model can produce. Drop a file onto the canvas and it is measured before it uploads, so the card is the right shape immediately.
Each flow remembers your last choice of model, ratio, duration, quality, and audio, and the next blank card in that flow starts there. It tells you when it has done this, with a label like "16:9 matching this flow". It never carries across to a different flow, and there is a Reset that goes back to the model's own defaults.
A preset that declares its own shape always wins. Picking Movie Poster inside a 16:9 flow still gives you a poster.
Running
Press Run for the whole flow, or run a single card, or run from a card onward. The price shows before anything starts, and it scales with quality, clip length, and the model.
Every card names its state:
Queued. Waiting for earlier cards. On most runs this is most of the cards.
Generating. Being made right now.
Rendering. The result is back and is being processed so you can play it.
Complete.
Failed.
The progress ring is measured against how long that model usually takes at those settings. It is an estimate, so it holds near the end rather than claiming to be done, and only completes when the result actually lands.
Assets on the canvas
Songs, voiceovers, videos, and images can sit on the canvas as their own cards. Drag files straight from your desktop, several at a time, or use the assets button on the canvas toolbar above the zoom controls to pull something from My Library in one click. That works during a run too.
A video card appears immediately and previews off your own disk while the file uploads, so you can move and connect it in the meantime. Run stays off until everything has finished uploading, and it tells you which file it is waiting on.
On a phone
The canvas works on a phone. You can open a flow, look at it, and work on it.