Show HN: wavecat – a fully local personal agent that watches your screen (wavecat.ai)

Samuel Yuan has unveiled wavecat, an innovative local personal agent designed to continuously monitor your screen, facilitating a deeper understanding of user activities while prioritizing privacy. This fully local application ensures that all data and processing remain securely on the user's device, eliminating reliance on cloud servers and safeguarding personal information. With a focus on personal AI, wavecat allows users to anticipate their needs without concerns about data breaches, making it a promising advancement for privacy-conscious individuals. Wavecat's installation includes language and vision model setup, requiring a substantial 19 GB of disk space due to the complexity of its billions of parameters. Optimal performance demands significant hardware capabilities, particularly for Mac users (24 GB of unified memory recommended) and Windows/Linux users with dedicated GPUs. Utilizing advanced technologies such as llama.cpp for inference and Qwen3.6 as its language model, wavecat aims for high efficiency, achieving comfortable processing speeds on powerful systems. Future updates are expected to enhance performance and expand the agent's functionality, including multi-language support and app integrations, heralding a new era for local AI applications.