The storyboard workspace

One side column plus a main area:

• Side column — top to bottom: Key elements, Audio, Text. All three are lists of what you have, so one narrow column is enough.
• Main area — Image and Video. You need to see these clearly, so the width goes here.
• Every border between columns can be dragged, horizontal and vertical alike, and the sizes you set are still there after a reload.
• The Image and Video columns are there even when empty, saying only that there is nothing yet — so a fresh canvas still shows you where things go. In the side column, Audio appears only once it holds something.

• Column name — which type this column lists.
• All ⌄ — filters this column: pick Generated or Not generated by status, or narrow it to one group on the canvas.
• Fill the view — top-right corner of the Image and Video columns. Click it and that column takes over the whole storyboard with its cards laid out as a grid; click again to fold it back. The side-column blocks do not have it — they are lists already, and spreading them out shows nothing more.

The Audio column holds tiles rather than cards: sound has no picture, so a note icon plus a duration is enough.

• Name — rename it in place from ⋯ → Rename, which is the same thing as renaming the node on the canvas (see 11.7).
• Picture — images show the finished image, videos a muted looping cover, text lays its body straight into the card. Anything not generated yet shows a placeholder and its status.
• Spec tags — which model, and the resolution or duration.
• Reference thumbnails — the row under the picture, showing what this one referenced (on the canvas, the edges running into it). Click one to jump to that reference.
• Hover a card and Chat and ⋯ appear to the right of the name: the first brings this piece into the chat in the right panel, the second gives you Set as key element / Rename / Make a copy / Delete.

Click a card and the details cover the main area from the right — the side column stays put on the left, so you can study a big image and keep clicking through the next ones of the same type. Piece by piece in 11.1.

• The detail header has Back, the name, ⋯, Add to chat and Close.
• Hover the picture and the toolbar floats up, exactly the same one you get when the node is selected on the canvas: Crop, Expression tuning, Multi-angle, Relight, Trim, Remove subtitles — all done right here, without going back to the canvas.
• Below the picture is the input bar: rewrite the prompt, switch models, regenerate.
• With the details open, the bottom list in the side column switches to the same type and its column name turns into a dropdown you can change the type with — so you can be looking at a video and go browse the dubs without closing anything.

The block at the top of the side column holds this film's people and places: what the lead looks like, where the story happens, which props turn up. In full in 11.4.

• Tag from the ⋯ menu on a card: choose Set as key element, then a category — Characters, Scenes, Props, Other. Tagged the wrong one? Same menu, Clear key element.
• The header of this block filters by category, so you can look at only characters or only scenes.
• It is not a second copy: the same image is here and still in the Image column.

(In full in 11.5 and 11.6.)

• Drag a card to reorder. That order is the same order the nodes have on the canvas, so both sides agree. You only sort within a column; dragging across columns does nothing.
• A card can also be dragged all the way onto the Agent input box in the right panel. Drop it and you get an @ reference, exactly the one you get from Chat on the card or from typing @ in the box yourself.

It filters names, prompts and body text as you type — one character in and the columns narrow together. On the Workflow side, search from Canvas elements in the left panel instead.