A simple voice assistant that respects your privacy. All voice assistants and smart speakers use proprietary systems and are spying on us 24/7. The goal of this project is to replace all of the spyware voice assistants in your digital life.
🟥 Confirmed not working yet
🟨 Partially working/in development
🟩 Fully working
🟦 Needs testing
⬛ Not implemented
Feature | Windows | MacOS | Linux |
---|---|---|---|
Speech recognition | 🟦 | 🟦 | 🟩 |
Text to speech | 🟦 | 🟦 | 🟩 |
Basic questions | 🟦 | 🟩 | 🟩 |
Advanced questions* | 🟦 | 🟦 | 🟩 |
Time | 🟦 | 🟩 | 🟩 |
Weather | 🟦 | 🟩 | 🟩 |
Smart device control | ⬛ | ⬛ | 🟨 |
Music control | ⬛ | ⬛ | ⬛ |
Nextcloud integration | ⬛ | ⬛ | ⬛ |
* Advanced questions are questions that require searching on the internet
Currently it is meant to be deployed on a desktop or laptop computer. In theory you can deploy it on a small computer like a raspberry pi and replace a smart speaker.
* Optional
** Optional but recommended
*** Optional but highly recommended
- Ollama
- Vosk model
- Internet connection
- IP to airport server
- Fast graphics card***
* Optional
** Advanced users only
- Install Ollama
- Download a Vosk model
- Put the location of the Vosk model in the
vosk_model_path
inconfig.py
- Set a custom wake word by changing
wake_word
inmain.py
* - Set a custom LLM model by changing
ai_model
inmain.py
** - Run
ollama serve
in a terminal window - Run the IP to airport server in a terminal window
- Connect a microphone and speaker to your computer (you can use a built-in one)
- Install the required python dependencies listed in the top of
main.py
- Follow the instructions in
personalization TEMPLATE.txt
- Enter the required API keys in
api_keys TEMPLATE.py
then rename it toapi_keys.py
- Customize
config.py
- Run
python main.py
in a new terminal window - Give microphone access to the program if necessary