Nodes, Nuclei, and How Your Graph Organises Itself

Every piece of content in your knowledge graph starts as a node — the atomic unit of knowledge in Ontoic. Nodes are individual, self-contained, and typed, but they are never truly isolated. From the moment you add one, Ontoic begins connecting it to every related node already in your graph. Over time, those connections accumulate, clusters form, and Ontoic crystallises them into something more: nuclei.

​

Nodes

A node is a single unit of knowledge. It can be a note you wrote, a PDF you uploaded, a web page you clipped, or a quote you extracted. Regardless of type, every node goes through the same enrichment process at ingestion:

• Typed — Ontoic records whether the node is a note, fragment, PDF, web clip, or quote
• Embedded — the node is described in the context of your existing graph and converted into a semantic embedding
• Connected — edges are drawn to related nodes based on meaning, not metadata

This means you never need to manually link two nodes. If a paragraph you wrote today echoes an idea in a PDF you added last year, Ontoic will place them near each other in the graph automatically.

​

Nuclei

When enough related nodes accumulate in one region of your graph, Ontoic crystallises them into a nucleus. A nucleus is an emergent summary of a topic your graph has developed — not a folder you created, not a tag you applied, but a pattern Ontoic detected in what you’ve been capturing. Nuclei form while you work, without any action on your part. Here is what that looks like in practice: a researcher building a knowledge graph on GLP-1 drugs might find Ontoic has automatically surfaced these nuclei:

• GLP-1 economics — 23 nodes
• Adherence & discontinuation — 17 nodes
• Policy landscape — 11 nodes
• Safety signals — 8 nodes

None of these were manually defined. They crystallised because the nodes were there, and the density of connections made the topic legible to the graph.

​

How nuclei help you

Nuclei act as a live table of contents for your thinking. Because they form from actual content rather than your intentions, they often surface subtopics you hadn’t consciously noticed were developing. A nucleus appearing in the corner of your graph can be a signal that you’ve been circling a question without realising it. You can query a nucleus directly using Ask, explore its constituent nodes, or use it as a starting point when you return to a research area after time away. Nuclei also give you an honest map of where your knowledge is dense and where the gaps are.