Initializing...
How it works: Each emoji has a description converted into a 512-dimensional
vector using feature hashing. Your query is converted the same way, then a WASM-powered
HNSW graph search finds the nearest emoji vectors in sub-millisecond time.
Everything runs locally in your browser — no server, no API calls, fully private.
Credits: WASM HNSW engine by
ruvnet/RuVector.