Subtitles and live captions on a Mac, without uploading the video
macOS Live Captions is free on Apple silicon Macs but cannot be exported. Veed's Lite plan is $19 a month. Rev charges from $1.99 a minute for human captions. Vocemo writes SRT and VTT files and captions system audio on your own Mac for $15 a month.
- macOS Live Captions
- Free, Apple silicon only, nothing to export
- Veed
- Lite $19 a month, or $108 a year, which is $9 a month
- Rev human captions
- From $1.99 a minute, 99% accuracy claimed
- The local option
- SRT and VTT files plus live captions, $15 a month
How do you get an SRT file from a video on a Mac?
Every route does the same two things. It turns speech into text, then splits that text into timed cues. The differences are where the video goes and how good the timings are.
- Upload it to a web editor. Veed, Submagic and Descript all generate subtitles after the file reaches their servers, then let you download an SRT.
- Pay a person. Rev charges from $1.99 a minute for human captions and quotes 99% accuracy. This is what compliance work and broadcast delivery usually needs.
- Run it on your Mac. A speech model on Apple silicon transcribes faster than the video plays and writes the SRT locally. Vocemo exports SRT and VTT and costs $15 a month.
Whichever you pick, read the result. Automatic subtitles get names, acronyms and product terms wrong, and those are exactly the words a viewer is looking for. Ten minutes of correction on a 20 minute video is normal.
SRT or VTT: which one do you need?
Both are plain text files listing a cue number, a start and end time and the caption text. The practical differences are small.
- SRT is the older format and the most widely accepted. Timestamps use a comma before the milliseconds. YouTube, Vimeo, Premiere, Final Cut and most players take it.
- VTT is the web format, required by the HTML video element's track tag. It starts with a WEBVTT line, uses a full stop before the milliseconds, and supports positioning and styling.
Rule of thumb: VTT if the video plays in a browser on your own page, SRT for everything else. Exporting both takes no longer than exporting one, so do that.
Does macOS already do live captions?
Yes, and it is worth using before you pay for anything.
Live Captions is in System Settings, then Accessibility, then Live Captions. It captions audio playing on the Mac in a floating window in real time. Apple's own documentation is clear on the limits: it is available only on Mac computers with Apple silicon, and it is not available in all languages, countries or regions. Apple also says the accuracy may vary and that it should not be relied on in high risk or emergency situations.
The bigger practical limit is that there is nothing to keep. Live Captions is a window of text that scrolls past. There is no transcript at the end, no file to save, no timestamps and no subtitle export. It is an accessibility aid for following along, not a way of producing a record. That is a reasonable design choice by Apple, and it is also the exact reason people go looking for something else.
How do you caption a video call or a stream that has no captions?
The audio is already available to capture. macOS has supported capturing system audio natively since macOS 13, so a tool can read what is coming out of your speakers without the virtual audio driver people used to install for this. If something tells you to install an audio routing extension on a current Mac, it has not been updated.
From there you have two shapes of answer.
- Read only, in the moment. macOS Live Captions. Free, no setup, nothing saved.
- Captions you keep. A local tool captions the system audio in a window and keeps the text, so a webinar, a training video with no subtitle track, or a foreign language stream ends with something you can search. Vocemo does live captions from system audio and exports SRT and VTT.
The confidentiality point matters here more than people expect. Captioning a customer call by uploading it to a web service sends the whole conversation to a third party. Captioning it on the machine does not. For the recorded version of the same problem, see screen recording with a transcript.
What do subtitle tools cost?
Prices checked September 2026, taken from each vendor's own pages.
| Tool | What you get | Price |
|---|---|---|
| macOS Live Captions | Live captions on screen, no export | Free, built in, Apple silicon only |
| Veed | Auto subtitles, styles, translation | Free: 2 minutes of auto subtitles a month. Lite $19 a month, or $108 a year. Pro $49 a month |
| Submagic | Animated burned in captions for social video | Free: 3 videos a month. Starter $19 a month, or $12 a month billed yearly |
| Rev | Human captions, 99% accuracy claimed | From $1.99 a minute. AI free for 45 minutes a month |
| Descript | Transcript editor, captions, video editing | Free: 60 minutes a month. Hobbyist $24 a month, or $16 billed annually |
| Vocemo | SRT and VTT export, live captions from system audio | $15 a month, or $144 a year |
If you would rather this ran on your own Mac, that is what Vocemo does. Free for 7 days, then $15 a month for 3 Macs.
Download free for MacRead the units before you compare the headline numbers. Submagic counts videos and caps their length. Veed and Descript count minutes or hours of media. Rev counts minutes of finished output. A local tool counts nothing, because your Mac is doing the work.
What does the local option not do?
Four gaps, and none of them are small if you happen to need that thing.
- No human accuracy guarantee. If you are captioning for a legal obligation or a broadcast delivery spec, an automatic transcript is a first draft. Rev sells the reviewed version, and that is what the guarantee is for.
- Sentence level timing on some models. Several speech models return no word by word timings, so captions fall back to sentence level timing spread across each piece. That is fine for a readable subtitle track. It is not enough for the word by word highlight effect that Submagic sells.
- No burned in styled captions. No animated social captions, no emoji, no per word colouring. A local tool gives you a subtitle file, and your editor burns it in if you want that.
- No shared workspace. Nothing sits on a server, so there is no review link, no comments and no team seats. That is the same property as the privacy, seen from the other side.
The requirements are also narrow: macOS 13 or later, Apple silicon only, and one licence covers 3 Macs. There is more on the general trade in running AI on your own Mac.
Which one fits which job?
- Following a meeting you cannot hear well. macOS Live Captions. Free, and already there.
- Captions for a course, an internal video or a customer call. Local. The content is usually confidential and an SRT file is all you need.
- Vertical social clips with animated captions. Submagic. That effect is what it is built for and a subtitle file is not a substitute.
- Accessibility compliance or broadcast delivery. Rev. Pay for the human pass and the accuracy claim that comes with it.
Questions
How do I add subtitles to a video on a Mac for free?
macOS has no built in subtitle generator, so the free routes are limited. Veed's free plan covers 2 minutes of auto subtitles a month and Descript's free plan covers 60 minutes of media a month, both after upload. For an unlimited local option you need a paid tool such as Vocemo at $15 a month.
Can macOS Live Captions be saved as a transcript?
No. Live Captions shows text in a floating window and keeps nothing. There is no transcript, no timestamps and no subtitle export. If you need a record you have to use a separate tool that writes the text to a file.
Does Live Captions work on an Intel Mac?
No. Apple's documentation says Live Captions is available only on Mac computers with Apple silicon. Apple also notes that it is not available in all languages, countries or regions, so check the macOS feature availability page for your region.
What is the difference between SRT and VTT subtitle files?
Both are plain text lists of timed caption cues. SRT is older and accepted almost everywhere, and puts a comma before the milliseconds. VTT starts with a WEBVTT line, uses a full stop before the milliseconds, and is the format the HTML video track tag requires. Export both if you can.
How much do captions cost per minute?
Rev charges from $1.99 a minute for human captions with a 99% accuracy claim. Automatic captioning is far cheaper per minute: Veed's Lite plan is $19 a month and Descript's Hobbyist plan is $24 a month. A local tool has no per minute cost at all.
Can I caption audio playing on my Mac without uploading it?
Yes. macOS has supported capturing system audio natively since macOS 13, so a speech model running on the Mac can caption a stream, a call or a video as it plays. macOS Live Captions does this for free without saving anything. Vocemo does it and keeps the text, at $15 a month.
Sources
Every price on this page was read from the vendor's own page on the date shown above. Prices change; if you find one out of date, tell us and we will correct it.
Competitor prices in this article were checked in September 2026. Prices change; if you find one out of date, write to us and we will correct it.