Working on the canvas
The canvas is an infinite, zoomable surface. Nodes snap to a 20-pixel grid, connections are drawn between them, and everything you do is saved immediately.
Navigation
| Action | How |
|---|---|
| Pan | Drag empty canvas, or two-finger scroll on a trackpad |
| Zoom | Mouse wheel / pinch. Range is 10% - 200% |
| Fit everything in view | Fit in the toolbar (the fit view button in the bottom-left controls in the agent layout) |
| Zoom in / out by step | The + / − buttons in the bottom-left controls |
| Overview | The minimap in the bottom-right shows the whole board with your viewport as a rectangle; drag it to jump |
Your position and zoom are remembered per board on this device. Below about 30% zoom connection lines are hidden for legibility (a small indicator tells you so); zoom back in to see them.
Creating nodes
There are five ways to put something on the board.
Toolbar. The toolbar across the top of the canvas (merged and board layouts) has a button for Notes, Events, Assessments, Embeds, Media, Locations, Documents, News Articles, Social Posts, RSS Feeds, Blockchain and Regions. A new node opens straight in the node sheet; Locations, Documents and Regions open a short form first. Actors, Organizations and Narratives are created in the Case file (same toolbar) because they are entities that can be shared by several nodes.
Keyboard. With the canvas focused, Shift plus a letter creates that type at the centre of your view - Shift+N Note, Shift+E Event and so on. Full table in Keyboard shortcuts.
Double-click empty canvas to drop a Note at that point.
Paste (Ctrl/Cmd+V) with the canvas focused:
| Clipboard contains | You get |
|---|---|
| An image | Media node, image copied into the board's storage |
| Twitter/X, Reddit or Bluesky post URL | Social Post node, post fetched immediately |
| YouTube URL | Embed node with preview |
| Direct image URL | Media node |
| Any other URL | News Article node - scraped and summarized in the background |
| Plain text | Note (first line becomes the title) |
Ask the assistant. Anything the AI can research it can also create. See AI assistant.
Selecting
- Click a node to select it. A bar with its title and an Open button appears at the bottom of the canvas (above the ask bar in the board layout).
- Shift+drag on empty canvas draws a selection box.
- Shift+click (or Ctrl/Cmd+click) adds to the selection.
- With several nodes selected the status bar shows the count and a Delete button.
- Esc clears the selection.
Selected nodes are also what the AI treats as "context" - the assistant sees whatever you have selected when you send a message. Selecting three articles and asking "what do these have in common?" is the fastest way to point it at specific material.
Editing
Double-click a node, or select it and press Enter or click Open, and it opens in the node sheet, a large panel over the board. Its right column always holds the node's tags, narratives and connections, Find an image (notes, events, locations and assessments) and Delete. The main column has two tabs:
- Overview - the node as a reader sees it: the text, the map for a location, the post for a social node, the wallet's balances, the profile for an actor, and so on.
- Edit - the fields. Every node type has its own: notes get a text area, events get a description, date, time, location and coordinates, media get an image, a caption and an Analyze button, and so on. A node you have just created opens here.
The title is editable at the top in both. Done, Esc or clicking outside the sheet keeps your changes; Save keeps them without closing, and Discard throws them away. Clicking one of its connections opens that node instead, keeping your edits to this one.
In the right column:
- Tags - free-form labels, searchable.
- Narratives - add the node to a narrative with a relevance, or remove it. Narrative nodes link to other narratives instead.
- Connections - the linked nodes; the pencil edits the connection itself.
- Delete - click twice; the first click asks for confirmation.
The date a node sits at on the timeline comes from the event date for Events, and otherwise from when the thing it describes happened (set by the assistant when it knows) or when the node was created.
Collapsing
Every node has a chevron in its header to collapse it to its title. The collapse / expand all button (Collapse in the toolbar; bottom-left controls in the agent layout) does the whole board at once, which is useful on dense boards before an auto-layout.
Node size
Text nodes (notes, events, assessments, articles, posts, web pages, people, organizations, locations, documents, wallets and feeds) size themselves to their text: a short note is a small card, and a longer one grows wider and taller, staying roughly square, up to about 540 × 675 px. Text beyond that fades out at the bottom of the card; the node sheet (double-click) always has all of it. Nodes grow with the Nodes text size in Settings → Appearance, so the same amount of text shows at any size. After nodes change size, Tidy spaces the board out again.
To give a node a size of its own, select it and drag its right edge, bottom edge or the grip in its bottom-right corner. The size is saved with the node. Right-click it and choose Fit to text to go back to automatic sizing.
Media and Region nodes have their own drag handles: media keeps its aspect ratio, regions are free-form.
Moving
Drag a node to move it. Drag any node in a multi-selection to move them all. Moving a Region moves the nodes inside it.
Deleting
Select and press Delete or Backspace, use Delete in the node sheet (click twice - the first click asks for confirmation), or the status-bar button for a selection. Deleting a node removes its connections. Deleting an Actor, Organization or Narrative node removes it from the canvas but keeps the entity - it is still listed in the Case file and can be placed again. To remove the entity itself, delete it there.
Undo and redo
Ctrl/Cmd+Z undoes and Ctrl/Cmd+Shift+Z (or Ctrl+Y) redoes node creation, deletion, moves and connection changes. The same controls are in the title bar. Text edits inside an editor use the editor's own undo.
Search
Press / for the search overlay. It matches against node titles on the current board as
you type; ↑/↓ and Enter jump the viewport to the node and select
it. For full-text search across node bodies, article text, actor bios and narratives, ask the
assistant - its search_board tool uses the database's full-text index.
Filtering what you see
- Filter by node type - the funnel in the bottom-left controls. Tick and untick types to hide them; counts per type are shown.
- Narrative filter - Narratives in the toolbar (a panel in the top-right corner in the agent layout), shown once the board has narratives. Choose a narrative and everything not in it is dimmed. See Narratives.
Filters affect only the display; hidden nodes are still on the board and still visible to the AI.
Right-click menu
Right-clicking a node offers Pivot…, which looks up the domains, IPs, emails, phone numbers, usernames or URLs the node holds and adds what they lead to (see Pivots). The rest of the menu is for narrative work: add to or remove from a narrative, ask the AI to suggest narratives for the node, or generate a summary of a narrative it belongs to.