“Hej Zeka” wake word model

Open-source openWakeWord model for the wake word “Hej Zeka”, trained on openwakeword.com from synthetic speech (Swedish). It runs fully on-device with the openWakeWord runtime: Home Assistant (openWakeWord add-on), Rhasspy, OpenVoiceOS and any Python project.

Model facts

Wake word: “Hej Zeka”
Language of the training voices: Swedish
Trained: 2026-08-29 · pipeline V4
Network: 64 × 3 layers, 80,000 training steps
Recall, clean speech: 95.2%
Recall, with background noise and room reverb: 79.8%
False activations on the benchmark set: 5.83 per hour at threshold 0.5
Files: ONNX (405 KB), TFLite (412 KB)
Downloads: 2

Recall is measured on 500 synthetic test utterances; false activations on about 10 hours of near-miss phrases, noise and read speech, one activation per frame at or above the threshold with a 2-second refractory period.

Download

Sign in on this page to download the model as ONNX (405 KB) or TFLite (412 KB). Downloads are free; models are licensed for personal and non-commercial use by default (commercial license).

Use it in Home Assistant

Download the TFLite file and copy it to the /share/openwakeword folder of your Home Assistant server.
Install the openWakeWord add-on and restart it so it picks up the new file.
In Settings → Voice assistants, open your assistant and choose “Hej Zeka” as the wake word.
Say “Hej Zeka” to your satellite or to the Home Assistant Voice Preview Edition.

Use it in Python

from openwakeword.model import Model

model = Model(wakeword_models=["Hej_Zeka_20260829_172441.onnx"], inference_framework="onnx")
# feed 16 kHz mono int16 audio in 1280-sample frames; a score above the threshold is an activation
scores = model.predict(frame)

Similar wake words in the library

“Hey guardian” — 58.4% recall
“Okay Kobold” — 82.2% recall
“Pirrko” — 65.6% recall
“Du Edgar” — 89.0% recall
“Hej Mira” — 85.8% recall
“Hej Tecke” — 92.0% recall

Custom wake words are trained in the Training Center; the library lists every public model with the same measurements.