Gisela is a personal voice assistant. Current skills:
- say useless phrases
- play music with different genres
- turn on and off lights
- Raspberry Pi Model 4b
- USB speaker and microphone
- optional: Sonoff RF bridge for IoT extenstions
- python min. 3.7
- picovoice engine
- optional MQTT for IoT
Go to console.picovoice.ai and setup an account.
In your user profile copy the access-key. Paste your access key in a file on project level and name it access-key.txt.
Train your wakeword (prefer to choose "Gisela") and place the training file (.ppn)
in folder /models. In controller.py rename the variable KEYWORD_PATH to the right filename.
Do the same with your intent training file (.rhn) and rename the CONTEXT_PATH variable.
Place your audio played when gisela did not understand you in ./audios/wakeword-callbacks.
Install required dependencies pip install -r requirements.txt.
Start your voice assistant with python3 picovoice-test.py.