Adding nodes
Every way to put a question, note, link, file or sketch on the canvas, from one-key hotkeys to paste and drag-and-drop.
When a question comes to you mid-read, or you find a source worth keeping, you want it on the canvas in a keystroke, next to the work it belongs with, before the thought slips away. Use whichever of these methods is closest to hand.
The toolbar and its hotkeys
The toolbar sits at the bottom center of the canvas. Click a node type or press its key, then click where you want the node. While you place it, the cursor is a crosshair and the toolbar reads "Click anywhere on the canvas to add" plus the node type. Press Esc, or the same key again, to cancel.
| Node | Key | Use it for |
|---|---|---|
| Chat node | C | Asking the AI a question |
| Text node | T | Your own notes |
| Image node | I | A picture from your computer (Add From Computer) |
| Scratch pad | S | A freehand sketch |
| Post node | V | A YouTube, TikTok, Instagram, Facebook, X or LinkedIn post |
| Document node | D | PDFs, Office and text files (a file picker opens when you click the canvas) |
| Web node | W | A web page (the toolbar calls it Web page node) |
| Orchestrator chat | A | A goal for the AI to build out on the canvas. It has no button: the chat opens mid-view, ready to type in. |
The toolbar also holds Tidy Up (0, see Organizing the canvas) and Search (⌘K, or Ctrl+K on Windows and Linux, see Search).
Note: The single-key hotkeys don't fire while you're typing in a chat composer, a Text node or any other text field. Click an empty spot on the canvas first.
Cmd-click for a quick chat
Hold ⌘ (Ctrl on Windows and Linux) and click any empty spot. A new chat node appears right there, next to whatever prompted the question. An empty canvas reminds you with "Use the Toolbar or Cmd/Ctrl + Click anywhere to create a node".
The right-click menu
Right-click an empty spot. Under Add, choose Chat Node, Text Node, Web Node, Post Node, Document Node or Image Node. The last two are labeled Beta.
The node appears where you right-clicked (Document Node opens a file picker first). Scratch pads aren't in this menu: use the toolbar or press S.
Drag from a node's handle
Drag from the handle on a node's right edge and release on empty canvas. You get a new chat node wired to the node you started from, so the chat sees it as context: the natural way to ask about a web page, a document or an earlier answer. See Connections and context.
From an Image node or a Scratch pad, a Create node menu opens instead, offering a connected Chat, Image or Scratch pad.
Double-clicking the right handle does the same without dragging (see Branching). New nodes only come out of the right handle: dragging from a left handle onto empty canvas shows "Nodes can only be created from the right handle".
Paste
Click the canvas and press ⌘V (Ctrl+V on Windows and Linux). Slashspace picks the most fitting node:
| You paste | You get |
|---|---|
| Images or other files | Image or Document nodes, by the drag-and-drop rules below |
| Nodes you copied | Copies of those nodes (see Organizing the canvas) |
| A YouTube, TikTok, Instagram, Facebook, X or LinkedIn post link | A Post node |
| Any other link | A Web node |
| Plain text | A Text node |
Pasted content lands at your mouse pointer, or mid-view if the mouse hasn't been over the canvas yet. As with hotkeys, pasting while you're typing in a text field puts the content there instead.
Drag files in from your computer
Drag files from Finder or File Explorer onto the canvas. Each file becomes a node where you drop it, and several files line up in a row.
- Images (PNG, JPG, GIF, WebP, SVG) become Image nodes.
- PDFs, Word, PowerPoint and Excel files, plus text and code files such as Markdown, CSV and JSON, become Document nodes.
- Anything else shows a message such as "File type .zip is not supported".
Nodes the AI adds for you
- Branches. Submit in new branch forks a conversation into a new connected chat node (Branching).
- Sub-agents. In Agent mode, the AI can hand parts of a task to new connected chat nodes (Sub-agents).
- Generated images. Each image from Image mode arrives as a new connected Image node (Image generation).
- Orchestrator. Give an Orchestrator chat a goal and it creates chat, text, web and post nodes and wires them together (Orchestrator mode).
Navigating the canvas
Pan and zoom around a large canvas, jump straight to any node, and choose how the canvas responds to your touchpad or mouse.
Connections and context
Control exactly what each chat reads by connecting nodes, choosing a context mode, and wiring in sources with mentions, the navigator and groups.