Graph databases answer "what's the shortest path between these two records" by walking index-free adjacency pointers directly, node to node, rather than joining tables. This simulator renders a random connected property graph in 3D and runs a real breadth-first search across it: a single expanding wavefront from the source, or two wavefronts — one from the source, one from the target — expanding toward each other until they meet. Live readouts track nodes visited and frontier size as the search progresses, and once a path is found the simulator silently runs the other strategy on the identical graph so you can see, in exact node counts, why bidirectional search is the standard approach for shortest-path queries in engines like Neo4j and Amazon Neptune.