My transcription tool of choice on the Mac, Mac Whisper, got another major update today with real-time meeting transcripts, significantly improved speaker recognition, tighter integration with Apple Intelligence, and a new transcription model. Here are the details.

Mac Whisper gets deeper Apple Intelligence integration. I’ve written in the past about how much I rely on Mac Whisper, particularly as part of my podcast production workflow. Since indie developer Jordi Bruin launched it in 2023, it has been on a relentless streak of improvements, including automatic meeting recording, speaker recognition, watched folders, system-wide dictation, and integrations with services like Notion, Zapier, Obsidian, and n8n.

Today, Mac Whisper was updated to version 15.2, introducing improved speaker recognition powered by NVIDIA’s Nemotron 3 Diarization model, supporting up to 8 speakers and live meetings in the new realtime meeting view. This feature allows users to follow, copy, or transfer a live transcript as the meeting unfolds, with recordings saved when the meeting ends. The update also adds Alibaba’s QWEN3-ASR transcription model with support for 22 languages, plus tighter integration with Apple Intelligence, Siri, Spotlight, and Shortcuts on macOS 27.

Users can now search transcripts system-wide, open them with Siri, pull text into Shortcuts, and refer to on-screen transcripts in Siri or Apple Intelligence requests. Mac Whisper 15.2 also adds folders and subfolders for organizing transcripts, with drag-and-drop support from the sidebar.