Knowledge graph

The knowledge graph is a force-directed visual map of your vault. Every note is a node. Every wikilink between notes is an edge. Every tag is a node too, and every note that uses a tag is connected to that tag’s node. The result is a picture of how your thinking is connected — clusters of related notes, hubs that link to many things, isolated notes that need attention.

The graph isn’t something you maintain. It’s something you watch grow. Every time you add a wikilink or a tag in notes, the graph updates. You don’t have to draw anything, label anything, or arrange anything. The connections come from the writing itself.

This page covers what the graph shows, how to navigate it, and the conventions that make it most useful.

What you’re looking at

When you open the Knowledge graph surface, you see:

• Nodes for notes (one per note) and for tags (one per tag)
• Edges for wikilinks between notes
• Sizes that reflect how connected a node is — a note with many incoming links is bigger than a note with one
• Colors that reflect the section, type, or theme of each note. The colors come from your theme so the visual language is consistent with the rest of the app.
• A force-directed layout where connected nodes pull toward each other and unconnected nodes push away. The graph settles into clusters naturally.

The layout is alive — it animates as the graph updates, and you can grab a node to drag it where you want it. The simulation will move things back toward a stable layout when you let go, but if you want to lock a node in place, you can pin it.

Navigating the graph

The graph is the primary navigator in the Knowledge graph surface, but it’s also the navigator on the Boards surface and in the docs page you’re reading right now. The same engine, the same conventions.

A few navigation moves:

• Click a node to open that note. The note opens in a panel next to the graph, so you don’t lose your place in the visualization.
• Double-click a node to navigate to the note full-screen.
• Drag a node to reposition it. Useful when you want to label a cluster or move a hub out of the way.
• Scroll to zoom. The graph supports deep zoom — you can pull in close to a single node or pull out to see the whole vault.
• Pan with the trackpad or middle mouse button. The canvas is infinite, so a large vault can sprawl.
• Search to focus. Type a note title and the graph highlights matching nodes and dims the rest. A fast way to find a note in a large vault.

A “fit to view” command resets the zoom and pan to show the whole graph, useful when you’ve gotten lost in a deep zoom.

Focus mode

The whole-vault view is great when your vault is small. When it grows, you want to focus.

Focus mode shows only the notes connected to a selected node — the “neighborhood” of that note. Pick a hub, and you see its immediate surroundings without the noise of the rest of the vault.

A typical use:

1. Open the graph and find a hub — a note with many incoming links, probably a project root or a topic index.
2. Enter focus mode on that hub.
3. The graph zooms to the hub and its neighbors. The rest of the vault fades out.
4. Click a neighbor to refocus on it. Step through the neighborhood one node at a time.
5. Exit focus mode to return to the whole-vault view.

Focus mode is also useful for context. When you’re reading a note, the backlinks panel shows the incoming links; the graph in focus mode shows the same connections visually, including the notes that don’t link back.

Filtering

The graph can be filtered by tag, note type, or folder. A filter narrows the graph to a subset of nodes and edges.

A few common filters:

• By tag. Show only notes with a given tag, plus the connections between them. The tag node itself is the natural center of the filtered graph.
• By folder. Show only notes inside a folder. Useful for separating work and personal contexts.
• By note type. Some setups use tags like #meeting, #idea, #reference to indicate the type of a note. Filtering by type shows the structure of just that type.
• By saved filter. Save a filter once and re-use it.

Filters can be combined. Show me the notes with #design that live in the Work/2026 folder, but not the ones tagged #archived. The graph updates immediately.

How the graph grows

The graph is a direct reflection of the wikilinks in your vault. Write a wikilink, add an edge. Remove a wikilink, remove an edge. The graph has no separate database to maintain.

A few conventions that make the graph most useful:

• Link liberally. The graph rewards links. A note with three wikilinks is a node with three edges. A note with twenty wikilinks is a hub.
• Tags are light edges. Tagging a note with #design adds a connection to the #design node. Over time, the tag nodes become visual clusters of related notes.
• Atomic notes make for interesting graphs. A note that holds one idea, expressed clearly, tends to be linked from many places. A note that holds a sprawling document tends to be linked from few. The graph is more useful when notes are atomic.
• The first paragraph of a note is its identity. Whatever you write at the top of a note shows up in the graph tooltip and the search results. Lead with the conclusion or the question.

The graph rewards a writing practice that emphasizes connections over containment. The act of writing is the act of building the graph.

Graph widgets on boards

A graph can also appear as a widget on a board. The widget is a live slice of the graph — drop a graph widget on a board, scope it to a tag or a folder, and it shows just the connections that matter for that board.

This is a useful pattern for project boards. The main board shows the project’s notes, tasks, and calendar; a graph widget in the corner shows how the project’s notes connect to the rest of the vault. As the project grows, the graph widget shows the growth.

The graph engine and this docs page

The same graph engine that visualizes your vault also powers the docs hub on this site. Each doc is a node; each wikilink between docs is an edge. The graph you see on the Boards docs page is generated by the same code that runs in your vault.

The reason this is worth mentioning: the docs page is a small example of what you can do with the graph on a board. The same patterns — nodes, edges, focus mode, filters, widgets — apply to your vault.

When the graph is most useful

The graph is most useful when:

• You’re starting a new project and want to see what existing notes might be relevant
• You’re trying to find an isolated note — a note with no incoming links is a candidate for retirement, or for being connected to its neighbors
• You’re exploring a topic — a cluster in the graph is a visual map of what you’ve written about it
• You’re writing — the graph shows you which notes might benefit from a new connection

The graph is less useful when:

• Your vault is brand new — there aren’t many connections yet. The graph fills in over time.
• You’re looking for one specific note — search is faster. Use the graph for the shape of the whole, not the location of a single thing.
• You’re trying to make a polished presentation — the graph is a thinking tool. For a talk, build a deck.

Tips and conventions

A few things that help the graph stay useful as your vault grows:

• Use folders for structure, tags for connections. Folders are good for top-level organization (Work / Personal / Archive). Tags are good for everything else. The graph rewards a tag-rich vault.
• Daily notes as hubs. A note named with today’s date is a natural place to dump the day’s thoughts and link out. The graph shows the daily note as a hub connecting everything that happened that day.
• Don’t over-link. A wikilink is meaningful. If everything links to everything, the graph is a hairball. Link where the connection is real.
• Review the orphans. Every so often, look for notes with no incoming links. They might be candidates for connecting to the rest of the vault — or for archiving.

See Notes & editing for the writing conventions that make the graph grow, and Boards for how to drop a graph widget on a board.