Demo broke us: Datastripes stalled on a 200MB sheet.

Fix = rip logic into WASM analytical DB (<3 duckDB), wire workers with synced message channels, cut the overhead. Parsing + aggregation now run parallel, no BS.

Benchmarks: 500MB ingested in ~2s Charts on 100k+ rows live Plain i7, 16GB RAM

Basically spreadsheets at GPU-speed, no GPU. Play with it: https://datastripes.com