While working on HyNote for Mac, I found it more useful to separate those decisions instead of placing one local label over the whole product.

1. Where transcription runs

A supported Mac can turn speech into text on-device once the required transcription model is ready. Initial setup can still require an account, an internet connection, and model downloads.

That tells you where the speech processing happens. It does not tell you where the resulting notebook will live.

2. Where the notebook is stored

Storage is selected separately. In HyNote, choosing On This Mac before recording or importing audio keeps that notebook on the Mac.

This distinction matters because “on-device transcription” and “local storage” answer different questions. A user should not have to infer one from the other.

3. Where AI summaries run

Summaries add a third path. Local AI has stricter requirements than the basic Mac app: Apple Silicon, macOS 26.4 or later, and Apple Intelligence ready on the device. Online AI is an optional paid alternative.

HyNote itself supports macOS 14 or later on both Intel and Apple Silicon. Putting the stricter requirement beside the feature it affects is clearer than presenting one compatibility statement for everything.

Capturing a meeting without a bot

The Mac app records microphone and system audio directly, so it does not add a participant to the call. System Audio permission is required, and the app captures audio rather than screen video.

Core recording and supported on-device transcription are free. The result is a live transcript and a notebook that can be stored in the location the user chooses.

The current build and setup requirements are here: HyNote for Mac

The wording is still something I am testing. When you see “on-device transcription” and “On This Mac,” do those labels make the processing and storage choices clear?