I recently came across Voice PE on YouTube and decided to give it a try. After some research, I found AmeriDroid offered the best shipping options, and soon, I had my new Voice PE unit. Setting up the hardware was straightforward, and I paired it with a pair of PC speakers for audio output. Initially, I configured it for full local processing, but I encountered issues with speech recognition accuracy and the absence of web assistant capabilities.
After some digging, I discovered that installing two Google integrations and obtaining a Generative AI API key could unlock Gemini 2.0 as my web assistant. This setup not only improved speech recognition but also allowed Gemini to control Home Assistant entities seamlessly. I’m now in the process of fine-tuning by adding aliases and exposing more entities to enhance functionality.
The transformation from a basic setup to a fully integrated smart home voice assistant has been incredible. Gemini now provides a natural and intuitive way to interact with my smart home devices, making daily tasks more efficient. I’m excited to continue refining this integration and exploring Gemini’s potential further. If anyone has tips or tricks for optimizing this setup, I’d love to hear them!