macOS has Live Captions already.
Here is what this adds.

Try the built-in one first

System Settings, Accessibility, Live Captions. It costs nothing, there is nothing to download, and the audio stays on your Mac the same way it does here. For a lot of people that is the end of the question, and this site would rather say so than pretend otherwise.

What follows is what the built-in one does not do. If none of it matters to you, keep your $9.

Where it stopped being enough

• The window is a window. Apple's captions sit in a panel you move and resize. This overlay takes no clicks at all: you work straight through it, and hold ⇧ only when you want to move it. On a call where you are also taking notes, that is the whole difference between glancing and rearranging.
• One model, no dial. Seven English checkpoints here trade latency against accuracy, from 160 ms upward: fast and rough for a conversation, slower and more careful for a lecture you cannot rewind.
• Language availability is not a setting. Apple's is not available in every language, country or region. The default model here detects what it is hearing across nine languages, and the menu bar switches between them.
• Everything, or one thing. Apple's transcribes audio from all apps by default. Point this at the tab playing the call and the rest of the machine stays quiet, which matters when music is playing somewhere else.
• Music does not become lyrics. A voice detector runs ahead of the recogniser, so a backing track never reaches it and does not fill the model's context.
• A new box when the speaker changes. Optional and off by default, because it runs a second model alongside the first.

Questions about the built-in one

Is Apple's Live Captions free?

Yes, entirely. It is part of macOS on Apple silicon Macs, under System Settings, Accessibility, Live Captions. If it covers what you need, there is nothing to buy here.

Why can I not find Live Captions on my Mac?

Apple's own documentation says the feature is not available in all languages, countries or regions, and it needs an Apple silicon Mac. If the panel is missing or greyed out, that is usually why.

Is this more accurate than Apple's?

There is no benchmark here that would let anyone answer that honestly, so this page does not claim it. What is measurable is the choice: seven English checkpoints trading latency against accuracy, where the built-in one gives you a single model and no dial.

Do I have to turn Apple's off to use this?

No. They are unrelated pieces of software and can run at once, though two sets of captions on one screen is mostly a way to compare them.

Does this do anything Apple's does not, on a call?

It stays on top and takes no clicks, so it survives switching to your notes; it can be pointed at one app instead of everything; and it can start a new caption box when the speaker changes.