Feature · Graph
See how everything connects.
Graph turns your memory into a living map. Explore it as a force-directed network of nodes, or as a relational view with semantic ribbons that name exactly why two memories are connected.
The short version.
Knowledge is not a list — it is a web. Graph makes that web visible. Nodes are sized by importance and coloured by type; clusters form around sources and themes; and ribbons between them are labelled with the relationship itself, like “Attribution” or “Inferred preference”. You can walk it hop by hop, or ask what changed since last week.
What you get.
- Force-directed canvas. An interactive graph where nodes are sized by importance and colour-coded by memory type.
- Relational ribbons. Sankey-style flows between clusters, labelled with the relationship — Attribution, Inferred preference, Lived context and more.
- Multi-hop exploration. Follow a thread from one idea to the next across the relationships that connect them.
- Graph Diff. Ask “what changed this week?” and watch new knowledge light up against the old.
- Dual view. Flip instantly between the graph and a sortable, filterable table of the same memories.
Under the hood.
From input to insight.
Questions, answered.
What do the ribbons between clusters mean?
Each ribbon is a named relationship between two groups of memories — for example “Attribution” (an entity to a fact) or “Inferred preference” (a fact to a preference). They show not just that memories connect, but why.
Can I edit a memory from the graph?
Yes. Click any node to open and edit that memory, or switch to the table view for sorting, filtering and bulk actions on the same data.