Hey everyone! I started building FUMI because I wanted to bring AI out of the browser and into a small physical companion you can actually interact with. The idea was to combine a Tamagotchi-style experience with real AI conversations, animated expressions, and a little life system that makes the device feel more alive. Getting the ESP32-S3, LVGL, audio pipeline, and AI backend to work together was quite a journey, especially with the hardware limitations. I wanted to make the project open source so other makers can explore it, build on it, and create their own versions. FUMI is still a work in progress, and I'd love to hear what you think. What would you add to a little AI companion like this? Website: https://fumi.codes/ GitHub: https://github.com/imaaiagent/fumi

FUMI

Build your own little AI companion.

Build your own little AI companion.