Keyword Detection (Live Audio + ML) — Zig

Source: apps/zig/heap/example_keyword_live/src/main.zig (also apps/zig/zeroheap/example_keyword_live/src/main.zig) | WASM demo not available — Zig 0.16 still lacks wasm32-emscripten support

Real-time 'yes'/'no' keyword detection using on-board microphones and the TensorFlow Lite Micro micro_speech model. Audio is captured via I2S DMA, spectral features are extracted, and a neural network classifier runs inference every ~1 second.

Language-specific patterns

This example uses the ove Zig module with allocator-aware constructors (Type.create(allocator)), exhaustive RTOS dispatch via switch (ove.target.current_rtos), narrow per-op error sets, defer-based cleanup, and std.log.* integration through ove.log.logFn.

See the overview for architecture details and the full API list.

How to build

# Native POSIX
make host.posix.example_keyword_live_zig
make configure && make download && make && make run