Homie
A desktop AI companion for spoken American English, growing toward a physical robot body.
- React / TypeScript interface wired to OpenAI, xAI / Grok and ElevenLabs for voice dialogue, persona and live subtitles.
- Local face detection and tracking in Python / OpenCV, mapped to on-screen eye direction; YOLO11n detection experiments.
- Perception → AI interaction → behaviour → control, designed as swappable modules.
- React
- TypeScript
- Python
- OpenCV
- YOLO11n
Head actuation and the quadruped form are in planning / prototype exploration — concept, not finished hardware.