OwlAIProject/Owl
An always-on wearable AI device that captures audio and visual data, processing it locally using Ollama-served LLMs like Llama2 and Mistral alongside Whisper for speech recognition.

Owl is a personal wearable AI system designed to continuously listen to and observe the wearer’s environment. It supports multiple capture devices including custom ESP32/NRF52840 hardware, Sony Spresense, and consumer devices like the Apple Watch. The system performs inference locally using Ollama with support for LLMs and VLMs, or optionally routes to commercial models. Audio processing uses Whisper for speech recognition with TTS support from Deepgram. It enables use cases like memory augmentation, proactive assistance, and distributed knowledge gathering.