Home / Sectors / In-browser analytical SQL

Analytical SQL.

BI-grade query on a million-row dataset, in the browser, at zero marginal cost.
The buyer's problem

Analysts iterate filters and group-bys constantly. Each refinement is a server query - latency, infra cost, and a moving rate limit. DuckDB-WASM solves the engine; the dispatch layer above it is the next bottleneck.

What Hadron replaces

The client-side aggregation layer that today is single-threaded JavaScript over DuckDB-WASM results - replaced with per-operator dispatch that keeps consecutive operators on the GPU.

What the buyer gets

A million-row analytical workspace that filters and pivots at interactive speed without round-tripping to a warehouse.

Patents demonstrated: P2 P3 Reference customers: in-browser analytical engines and BI platforms
How to tell Hadron is working

Hadron is not claiming the GPU is fastest. It picks the right backend per operator - the same SQL query may route filter to CPU and groupBy to GPU. Watch the Dispatch evidence panel beside each app: every preset and predicate change prints the per-operator plan, the score that drove each routing decision, and the rows in / out.

GPUHardware-accelerated - chosen when batch size, shape and precision all favour it.
WORKERBackground thread - chosen when GPU transfer cost would dominate.
CPUInline - chosen for small workloads where dispatch overhead would dominate.
REFUSEDGPU categorically refused (patent-protected) - specific predicates that would lose on GPU.
UI heartbeat: 0 frames painted Longest frame gap (last 5s): 0 ms main thread responsive
How to read this page Three applications below: a taxi explorer over 1M rides, a cohort builder with live size feedback, and an A/B test analyser. Every chart and table you see is built in your browser tab with no server.
Loading Hadron query + compute bundles...