Which operating systems does Um Out support?

macOS 13 Ventura or newer on Apple silicon (M1 or later); Intel Macs are not supported. On Windows, Windows 10 or 11, 64-bit.

How large is the Um Out speech model download?

The default model is roughly 1.4 GB. It is not included in the installer and is downloaded once, the first time you run a tool that reads what was said. Tools that do not read speech, such as the quiet gap remover, work without it.

← Help Center

Getting started

Um Out system requirements

Q: What version of Premiere Pro do I need for Um Out?

Premiere Pro 25.6 or newer. This is enforced by Adobe, not a recommendation: the plugin declares 25.6 as its minimum, and Premiere will refuse to install it on anything older. There is no upper limit, so newer releases are supported as they arrive.

Q: Does Um Out need an internet connection?

Yes, for activating your licence, downloading the speech model the first time you transcribe, and any AI feature you have explicitly turned on. Your footage is not one of them: analysis runs on your own machine, and video and audio are not uploaded.

Q: What languages does Um Out support?

English. Other languages are not supported today.

Um Out is a panel inside Adobe Premiere Pro, so Premiere sets most of the bar. The parts that are specific to Um Out are the Premiere version, Apple silicon on Mac, and a one-time speech model download.

What version of Premiere Pro do I need?

Premiere Pro 25.6 or newer. This is enforced by Adobe, not a recommendation: the plugin declares 25.6 as its minimum, and Premiere will refuse to install it on anything older. There is no upper limit, so newer releases are supported as they arrive.

Which operating systems are supported?

macOS 13 Ventura or newer, on Apple silicon (M1 or later). Intel Macs are not supported.
Windows 10 or 11, 64-bit.

If Premiere Pro 25.6 already runs well on your machine, it meets the hardware Um Out needs to open and edit your timeline. The exception is transcription, below.

Do I need an internet connection?

Yes, for three things: activating your licence, downloading the speech model the first time you transcribe, and any AI feature you have explicitly turned on.

Your footage is not one of them. Analysis runs on your own machine, and video and audio are not uploaded.

What is the speech model download?

The tools that read what was said — filler words, captions, best take, stream highlights — need a speech recognition model. It is not included in the installer, because it is far larger than the installer itself. The first time you run one of those tools, Um Out downloads it once and reuses it afterwards.

The default model is roughly 1.4 GB. Budget the disk space and expect the first transcription to take longer than later ones. Tools that do not read speech, such as the quiet gap remover, do not need it and work without the download.

What languages are supported?

English. Other languages are not supported today.

How many computers can I use one licence on?

Two. You can move a licence between machines from the account portal.

Do I need my own AI API key?

No. Um Out's core tools run without one. If you would rather use your own provider account for the AI features, see setting up your own API key.

Ready to install? Follow the install steps, or start with the 7-day free trial.