I am building a custom voice assistant that uses voice detection every couple of seconds to decipher what im saying then from scratch that will be able to do basic calculations, check the weather, control my room and using a gemini API be able to directly answer more advanced questions. I also plan to put in a mini PI and add a screen with animations and have it sit on my desk
Today I debugged some of my code, but along with that, I redid all of my weather features. Instead of only getting the temperature, you can individually ask for the temperature, rain forecast, and cloud coverage. If you ask for the weather, it automatically gives you all three.
Now i have added feature where in order to activate the assistance you must say Hey Jarvis, it uses the pre trained openwakeworld model for detection.
Over the past few hours, I have created the base code for it to perform calculations, check the weather, and use a Gemini API to answer questions. It then uses the built-in voices from your laptop to speak the answers out loud.
I plan to add memory, timers, and stopwatches, along with a small physical cube powered by a Raspberry Pi with a screen for animations. It will also have a feature where you can activate it by simply saying “Hey” followed by its chosen name.
All of the voice recognition is handled by Google’s AI, so it can run on any laptop or Raspberry Pi. Finally i also plan to add more weather features including future forecast and precipitation.
