Knowledge Graph
See how your research, projects, decisions, and ideas connect. MindWiki's knowledge graph gives you another way to find useful context, identify well-connected pages, and notice material that deserves a clearer place in your knowledge.
Pages appear as nodes. Relationships appear as edges. Some edges come from links you wrote; others express structural associations such as shared tags or folders.
Open the graph
In the Mac app, open the command palette with Cmd+K and choose Knowledge Graph, or use the available View menu action.
The visual canvas is a Mac feature. Graph data is also available to authorized agents through mindwiki_graph and through the REST API. Access to graph data does not mean the web app has the same canvas interface.
Understand the relationships
| Relationship | What it means |
|---|---|
| Wikilink | One page explicitly references another. |
| Shared tags | Pages use one or more of the same topic labels. |
| Title similarity | Titles share a naming signal. |
| Same folder | Pages have a common file location. |
| Same area | Pages share an area property. |
The layout gives stronger weight to explicit references than to broad organizational proximity. Connected pages influence one another's positions in the force-directed map.
The reason for an edge matters more than its appearance. Two pages in the same folder are associated by organization; that does not mean one supports the other's claims. A similar title is a clue to investigate, not a confirmed semantic relationship.
Read useful patterns
Hubs are highly connected pages. A project index, topic overview, or recurring reference can become a useful entry point because many pages point to it.
Clusters are groups of connected pages. They may represent a project or theme, but can also reflect common tags or folder structure. Open the pages before assigning meaning to the shape.
Bridges connect groups of material. A source linked from two projects may reveal a useful shared dependency or idea.
Orphaned pages have no explicit incoming or outgoing wikilinks. An orphan is a candidate for review, not automatically a bad note. It may still appear near other pages through inferred structural edges.
Navigate the canvas on Mac
- Click a node to open its page.
- Drag a node to reposition it while exploring.
- Scroll or pinch to zoom.
- Drag empty space to pan.
The graph uses cached data to make reopening responsive. Recent changes can take a save, sync, or refresh cycle to appear.
Use the graph to improve a project
Start with one question: “Where should I begin reading about this project?” Open its hub, inspect the linked sources, and check whether key decisions are connected.
When you find a useful missing reference, add a wikilink in the sentence where it makes sense. Add shared tags only when the pages genuinely share a theme.
The goal is clearer navigation and understanding. You do not need to maximize the number of edges or make every note part of a dense cluster.
Ask an agent to inspect graph structure
Try:
Inspect my MindWiki graph for the Website Launch project. Identify useful hub pages, explain the relationships you used, and read the relevant pages before recommending new links.Or:
Find pages without explicit incoming or outgoing links. Suggest a relevant project or source connection for the ones that would benefit from it.Graph tools help locate candidate context. Reading the source pages is still necessary before drawing conclusions or proposing changes.
Access graph data through the API
curl -H "Authorization: Bearer $MINDWIKI_API_KEY" \
'https://api.mindwiki.io/vault/graph'This requires an API key with read access. Keep keys out of page content and AI conversations. Use the API reference for the current response shape and limits.
Next: Wikilinks & Backlinks, Vault Health & Permissions, and Search & Views.