How to Add Subtitles to a Screen Recording on Mac (On-Device)

macOS doesn’t have a built-in button that adds subtitles to a screen recording. To add subtitles to a screen recording on a Mac you have three routes: write a subtitle file by hand, upload the audio to a cloud transcription service, or transcribe it on the Mac itself with Final Cut Pro or a tool built on a local speech model. The third keeps your voice on your computer.

Below we compare all three, show what a subtitle file looks like, and explain how to check the result before your client sees it. If you’re still setting up the recording, start with how to screen record on Mac.

Why subtitles are worth the effort

• People watch on mute. Your client opens the link on a phone between two meetings, sound off. Without subtitles, a walkthrough is a silent screen.
• Accessibility. The W3C’s Web Content Accessibility Guidelines (WCAG 2.2) list captions for prerecorded video as a Level A success criterion, the most basic level. If a client has hearing loss, subtitles are the difference between understanding you or not.
• Non-native speakers. Many agencies work across languages. Reading along helps a client who follows spoken English, or Italian, more slowly than written.
• Scanning. Subtitles let someone find the part about the pricing table without rewatching the whole thing.

We go deeper into the client side in subtitles get your videos watched on mute.

What a subtitle file looks like

Subtitles are usually a separate text file next to the video. The two formats you’ll meet most:

• SRT: numbered blocks, each with a start and end time and a line of text. Very widely supported; Final Cut Pro, for example, lists it among its caption formats.
• WebVTT (.vtt): the W3C format for text tracks in web video. It starts with the word WEBVTT, then timed cues.

A WebVTT cue looks like this:

WEBVTT

00:00:01.000 --> 00:00:04.000
Here's round two of the homepage.

00:00:04.000 --> 00:00:07.500
The main change is the hero section.

The alternative to a separate file is burned-in subtitles, drawn into the video picture by an editor.

	Separate file (SRT, VTT)	Burned-in
Viewer can turn them off	✓	✗ always on
Shows on any player	~ depends on the player	✓
Fix a typo	✓ edit the text file	✗ export the video again

Option 1: write them by hand

For a 60-second video, typing an SRT or VTT file in a plain-text editor is realistic. Play the video, note the times, write the lines. Past two or three minutes it becomes tedious, and it’s easy to drift out of sync. It’s private and free, but rarely worth your time.

Option 2: a cloud transcription service

Many web services and video platforms transcribe your audio on their servers and give you a subtitle file back. They’re fast and need no setup.

The trade-off is privacy. To transcribe, the service receives your voice and everything you say: client names, prices, the bug in their checkout. Before using one for client work, check where the audio is processed and stored, how long it’s kept, whether it’s used to train models, and whether the provider offers a data processing agreement. Our article on on-device vs cloud transcription goes through these questions, and GDPR and client videos covers who is responsible for the data.

Option 3: transcribe on the Mac

Speech recognition models now run well on a Mac, so you can create subtitles without uploading anything.

Final Cut Pro

Apple’s professional editor (a paid app) can create captions automatically. Select the clips in the timeline and choose EditGenerate CaptionsClosed Captions, or press ⇧⌘C. Apple’s guide says the feature requires a Mac with Apple silicon and is available in U.S. English only. The first time, Final Cut Pro downloads the language model, which needs an internet connection once. Automatically generated captions are in iTT format.

It’s a good option if you already use Final Cut Pro and speak English. It’s a lot of app for a three-minute client update.

Tools built on Whisper

Whisper is an open-source speech recognition model from OpenAI, released under the MIT License. It handles many languages and runs locally. Its own command-line tool can write SRT and VTT files directly, and several Mac apps wrap it in a simpler interface. Setup ranges from “drag in a file” to installing Python. We describe what it’s like to run on a Mac in running Whisper on a Mac.

Apple’s speech frameworks

For developers, Apple’s Speech framework recognizes speech in recorded or live audio. A request can be set to keep audio on the device, and macOS 26 added a newer API, SpeechAnalyzer, with a SpeechTranscriber module for general transcription. These are building blocks for apps, not something you use directly.

Option	Voice stays on your Mac	Effort	Languages
Write by hand	✓ nothing is sent	High	Any you can type
Cloud service	✗ sent to the provider’s servers	Low	Depends on the service
Final Cut Pro	✓ after a one-time model download	Medium, paid app	U.S. English only
Whisper-based tools	✓	Low to medium	Many

Check the subtitles before you send

No speech model is perfect. Apple’s documentation for its Speech framework notes that on-device requests won’t be as accurate as server-based ones. Mistakes cluster in predictable places.

Clear audio gives any model less to guess: a microphone close to your mouth and a quiet room help. See mic setup for client videos. If your clients care where their data goes, our GDPR-friendly screen recorder page explains how JustRec handles it.