Always-on keyword spotting that listens for a specific activation phrase. Ultra-low-power operation is critical since the model runs continuously waiting for the trigger word. Uses DS-CNN or depthwise separable convolutions on mel-spectrograms of 1-second audio windows. The model classifies short audio segments as either the target keyword or background noise. Common in smart speakers, voice-activated appliances, and hands-free industrial interfaces.
| Minimum RAM | 64 KB |
| Minimum Flash | 512 KB |
| Sensor Inputs | microphone |
| Typical Model Size | 40 KB (quantized int8) |
| Minimum Clock | 48 MHz |
No hardware guides for wake word detection yet. Use the MCU Checker to find compatible hardware.
ForestHub compiles visual AI workflows to C code for your microcontroller. Choose your hardware, build your wake word detection pipeline, deploy in minutes.