Vector Search & HNSW — pick a scenario
0/3 donePick one and press Start, then click Next › to step through it.
Watch forThe path taking long jumps on the top layer, then short steps once it drops to layer 0.
More about this scenario
You'll learnHow a layered graph finds the nearest vector in about log(n) hops instead of scanning all n.
How it runsEnter at the top entry point, hop to the closest neighbour until none is closer, drop a layer, and repeat down to layer 0.