This project is the server-side proxy component of the Voice Fill Test Pilot experiment. Roughly, it's the box labeled "speak-to-me node server" in the diagram below.

docker build -t asr-server .

docker run -e ASR_URL=http://192.168.0.44/asr -e DISABLE_DECODE_JAIL=1 -p 9001:9001 speech-proxy:latest