.svg){kind=small}
The Lab.
Reports, experiments, demos and systems that make our real-time AI research inspectable and fundamentally measurable.
02 / Lab Work
LAB / 001 · BASELINE A1ESTABLISHED
Instrumenting the real-time audio path.
A controlled baseline measuring browser capture, WebM/Opus chunking, transport and server processing before introducing speech or model inference.
98.0 msBest observed client round-trip
18.7 msBest observed server processing
48 kHz · Mono · WebM/Opus · ~8.3 chunks/sOPEN TECHNICAL REPORT→
LAB / 002 — CONCURRENT EXECUTIONIN PROGRESS
Keeping interaction alive while actions execute.
We’re testing whether speech, state updates and tool execution can run concurrently instead of forcing the system into a sequential listen → reason → act → wait → respond pipeline. The goal is to keep the interaction responsive while work continues in the background, without losing state or execution context and we are in working we didnt invent numbers yet
Concurrent pipeline · Background execution · Stateful context