Explore the semantic connections inside an MMPM memory substrate as a force-directed 3D graph. Each node is a memory atom; each link is a typed edge or a Markov arc — the learned probability that recalling one memory leads to another. Search to seed the graph from a concept, then click any node to expand its neighbourhood and follow how knowledge connects.
The knowledge graph is how MMPM turns a flat store of atoms into reasoning structure: edges capture relationships (supersedes, depends-on, derived-from, member-of) while Markov arcs capture usage patterns learned over time. The explorer runs in your browser with WebGL. To learn how atoms, edges, and arcs fit together, see the documentation.