Runs Rhasspy services using supervisord. Service configuration is automatically generated using rhasspy-supervisor.
- Audio Recording
- Audio Playback
- Wake Word Detection
- Speech to Text (ASR)
- rhasspy-asr-kaldi-hermes built on rhasspy-asr-kaldi
- rhasspy-asr-pocketsphinx-hermes built on rhasspy-asr-pocketsphinx
- Built on rhasspy-asr library (pypi)
- Uses rhasspy-silence library (pypi)
- Intent Recognition (NLU)
- rhasspy-fuzzywuzzy-hermes built on (rhasspy-fuzzywuzzy)
- rhasspy-nlu-hermes built on (rhasspy-nlu)
- Text to Speech (TTS)
- Dialogue Management
- Remote/Command
Before building Rhasspy, you will need some support packages:
python3*(Python)git(clone repo)build-essential(compiler)libatlas-base-dev(snowboy)swig(snowboy/pocketsphinx)portaudio19-dev(pocketsphinx)
sudo apt-get update
sudo apt-get install \
python3 python3-dev python3-setuptools python3-pip python3-venv \
git build-essential libatlas-base-dev swig portaudio19-devInstall runtime dependencies:
supervisord(process management)mosquitto(MQTT broker)sox(WAV conversion)alsa-utils(record/play audio)libgfortran4(Kaldi)- Text to speech
espeakflitelibttspico-utils
- Miscellaneous
perlcurlpatchelfca-certificates
sudo apt-get update
sudo apt-get install \
supervisor mosquitto sox alsa-utils libgfortran4 \
libfst-tools libngram-tools \
espeak flite libttspico-utils \
perl curl patchelf ca-certificatesIf you can an error regarding libttspico-utils, you can skip installing it.
The necessary .deb files can be manually downloaded and installed from http://archive.raspberrypi.org/debian/pool/main/s/svox/. You will need libttspico-utils and libttspico0 packages with matching versions.
Clone the repo and build:
git clone --recurse-submodules https://github.com/rhasspy/rhasspy-voltron
cd rhasspy-voltron/
makeRun the bin/rhasspy-voltron script with a --profile <LANGUAGE> argument.