vosk

Vosk & Kaldi

Run server:

podman run -ti --rm -p 2700:2700 docker.io/alphacep/kaldi-en:latest

Then run client:

./test_microphone.py

Pros:

• No direct wake-word functionality
• Ready to run container with Websocket integration

Cons:

• Need to filter out "wake word" variants ( "he wrote me" -> "hey rodney" )
• Accuracy is poor