Show HN: Minute – Offline meeting notes on macOS with Whisper and llama.cpp (github.com)

Minute, a newly introduced offline meeting notetaker for macOS, leverages the powerful whisper and llama.cpp technologies to record, transcribe, and summarize meetings without any reliance on cloud services. This innovative app operates entirely on your device, providing real-time transcription and the ability to generate meeting summaries, decisions, and action items by utilizing local machine learning models. Users can record audio, which is transcribed live, and even interact with their notes via plain-language queries—all while ensuring privacy and security, as the app never connects to the internet after its initial setup. This development is significant for the AI/ML community as it showcases a strong shift toward privacy-centric applications that can function independently of cloud infrastructure. By combining on-device processing with state-of-the-art transcription and summarization capabilities, Minute addresses essential concerns regarding data security while maintaining a high degree of usability. Technical aspects include real-time speaker detection, customizable model selection based on hardware capabilities, and a structured file system for easy note management, further solidifying Minute’s role as a pivotal tool in the realm of AI-driven productivity software.