hardware · keiaudio

KeiAudio — audio co-processor concept.

A concept co-processor design for real-time audio pipelines. Built to run signal-processing and inference inline, at low power, without round-tripping to a general-purpose CPU or GPU — no fabricated silicon yet.

What it does

A planned dedicated chip that would take raw audio in and emit processed audio plus inference results out. Everything stays on-device by design — no streaming to a server, no wake-word leaking to a cloud.

Why dedicated silicon

General-purpose CPUs and GPUs leave power on the table for steady-state audio. A fixed pipeline on dedicated silicon hits the latency and energy budget that battery-class devices actually need.

Roadmap target

Dedicated silicon or an FPGA fabric implementation, in that order of preference — the lab already runs hardware accelerators in an FPGA fabric elsewhere (see the MAX-OS silicon page), which is the natural proving ground before any fixed silicon commitment.