Skip to content

Fact · Dictation software

11 = yes, 0 = no

Local Whisper's security page says that after the model download, inference stays in process on the Mac.

Quote found word for word on the live source, last checked October 25, 2026

In the source's own words

“After the selected 1.7B or 0.6B snapshot is downloaded, inference remains in process on the Mac.”

Local Whisper privacy and security reference, Local Whisper (gabrimatic), 2026, Runtime boundary.

Open the original
What it measures
The sentence is about the Qwen3-ASR models. The runtime table on the same page lists Parakeet-TDT v3 as in-process MLX too.
What it does not show
Setup and model downloads use the network, as the page says. The sentence speaks of the Qwen models only.

In the data

A row of Dictation App Features Compared: Offline, Privacy, AI, as Local Whisper: speech on device: 1 1 = yes, 0 = no.

App
Local Whisper
Feature
Speech processed on the device
Detail
in-process MLX on the Mac
As of
2026-10
Note
1 means the maker's page says the app does this.

Source