HyNote for Mac: a walkthrough of local meeting transcription and notes

Hi everyone, I’m Ryan, and I work on HyNote. Here’s a short walkthrough of the Mac app for anyone exploring on-device speech transcription and meeting notes.

The setting to check first is storage. Choosing on-device transcription controls where speech is processed; it doesn’t make a cloud note local. To keep the note on your computer, choose On This Mac in New Note before recording.

1. Capture both sides of the conversation

After signing in and allowing the recording permissions, choose Record → Record Meeting. This captures your microphone and audio from meeting apps running on the Mac, without adding a bot to the call. Voice recording captures only the microphone.

macOS may call the permission “Screen & System Audio Recording.” HyNote uses it for audio, not screen video.

2. Check the transcription model before starting

Select on-device transcription and the language being spoken. Wait for the language model to be ready, then make a short test recording and check that the waveform and live text respond. Available languages depend on your Mac and macOS version.

3. Review the transcript alongside a summary

Finish the recording, wait for processing, then open Notes to create a summary or action-items card. Chat lets you ask follow-up questions about the note. Check assignments and decisions against the transcript before sharing them.

These screenshots are interface examples from our guide, not an accuracy benchmark or one continuous recording.

A few requirements before trying it: the app needs macOS 14+ and sign-in. Local AI has stricter requirements: Apple silicon, macOS 26.4+, and Apple Intelligence enabled with supported language/region settings and models ready. Working on-device transcription doesn’t necessarily mean Local AI is available. Recording and supported on-device transcription are free; Online AI is optional and paid.

Download and Mac requirements · Full walkthrough

For people building similar apps: how do you show the difference between local processing and local storage before someone starts recording? I’d be interested in examples that make both choices clear without another settings screen.