Meetily: A Trending Open-Source, Privacy-First Local AI Meeting Assistant Built in Rust
The open-source repository Zackriya-Solutions/meetily has gained significant traction, securing over 13,000 GitHub stars this month. The tool offers 100% local AI meeting transcription, speaker diarization, and summarization for macOS and Windows. As privacy concerns grow around cloud-based AI services, Meetily provides a secure, self-hosted alternative that processes sensitive meeting data entirely on-device. Its use of Rust and optimized models like Parakeet and Whisper ensures high-speed performance without relying on external APIs. Meetily achieves four times faster live transcription by leveraging Parakeet and Whisper models, and integrates Ollama for local summarization. It supports speaker diarization to identify who spoke when, and is designed to run without any cloud connection.
## BACKGROUND
Speaker diarization is the process of partitioning an audio stream into segments based on speaker identity, answering the question of "who spoke when." Ollama is an open-source platform that allows users to run large language models locally on their computers. NVIDIA's Parakeet models are designed for highly efficient and accurate automatic speech recognition (ASR).