Recognise through the server or on the device?

Applies to: Windows, macOS, Linux

Stenowise recognises your dictation either through the Stenowise server or directly on your computer. “Through Stenowise” is the default setting.

The difference

	Through Stenowise	On this device
Where recognition runs	on servers in the EU	on your computer
Recording leaves the computer	yes	no
What the computer needs	no special requirements	memory and, depending on the recognition tier, graphics

Through Stenowise, the recording and your dictionary go to the server, the recording already while you speak. On this device, the recording becomes text on your computer and is then dropped. With text cleanup on, the recognised text goes to the Stenowise server in the EU for text cleanup. For the exact data that leaves your computer, see Which data leaves my computer?

When recognition on the device makes sense

The recording must not leave your computer.
You dictate without a connection. With text cleanup on, Stenowise then inserts the raw version.

With text cleanup off, no dictation content leaves the computer during recognition on the device. For the weekly limit, Stenowise reports only the word count.

Change the recognition

Open Settings › Privacy.
Under Recognition, choose “Through Stenowise” or “On this device”.

You can also change the recognition under Recognition in the menu of the icon in the notification area of the taskbar (Windows), the menu bar (macOS) or the notification area (Linux).

Recognition on the device needs a recognition tier that you download once. Until a tier is downloaded, Stenowise recognises through the server. To download a tier, see Download, switch and delete recognition tiers.

Frequently asked questions

Yes, Stenowise needs a sign-in in both cases. An existing sign-in also works without a connection.

Do dictations on the device count towards the weekly limit?

Yes. Stenowise reports the word count for them as soon as a connection is available again.

Last updated 28 September 2026