How to transcribe video to text on a Mac, without uploading it

Does Mac have built-in transcription?

Yes, for speech you record or play live, in three places:

Apple's Notes guide lists English, Spanish, Portuguese, Italian, French, German, Japanese, Korean and Chinese for audio transcription, and Live Captions and Voice Memos transcription are not available in every country or region. What none of the three guides describes is taking in a finished video and handing back a transcript with timings, or a caption file. For that you need an app made for video.

Does QuickTime have transcription?

No. QuickTime Player's user guide covers playing, recording, trimming and exporting, and it can show subtitles or captions a file already carries, but it has no transcription. You can play the video with Live Captions on and read along, but Apple's guide describes no way to save what it shows, let alone with timings.

How do you transcribe an already recorded video on a Mac?

With Sharpy, from any recording to text with timings:

1. Drop the recording on the window. Sharpy makes an edit from it, picture and sound linked, ready for you or the agent.
2. Name the language, or let Sharpy hear it. It transcribes in the language you name, or the one it hears where speech begins, and says how sure it is. When it is less than half sure, it asks you. A recording with no speech gets no transcript, and Sharpy says so rather than inventing words.
3. Let it transcribe. Two transcription engines run on your Mac and vote on every word, then each word is timed to the sound.
4. Check the underlined words. In the WORDS tab, a word the engines were unsure of is underlined, so you check a handful instead of reading every line.
5. Correct what it heard wrong. Right-click, choose Correct, and type what was said.
6. Take the text where you need it. As SRT and WebVTT files, a subtitle track inside the video, or the transcript itself through the agent.

For languages written two ways, such as Hindi and Urdu, or Serbian, Croatian and Bosnian, name the writing once: Sharpy keeps it with your dictionary and uses it for every later recording of that speech.

How accurate is transcription on a Mac?

Accurate enough that checking takes minutes, not hours. On Google's FLEURS test recordings, 20 speakers a language, Sharpy got 3.56% of words wrong in German, 3.81% in Spanish, 4.45% in English and 5.35% in French. For 25 European languages, both engines vote on every word; in other languages one engine's words make the transcript, and Sharpy says so.

It is quick, too. On the base Mac mini M4 with 16 GB, the whole transcript pass takes 4.9 minutes for an hour of sound, and recordings up to four hours long are transcribed and split by speaker on a 16 GB Mac. Nothing is uploaded, and there is no limit on minutes.

Who said what comes with it. Sharpy splits a recording into speaker turns, so you can, for example, find every question the interviewer asked. Tell it how many people speak and the count is exact; left to count alone, its weak point is a speaker who says very little.

How do you fix words it heard wrong?

Names, brands and jargon are where any transcription slips. Sharpy gives you three ways to fix them, from one word to every recording you make:

• Correct a word. In the WORDS view, select the words, choose Correct, and type what was said. It can be more or fewer words than were heard. The new words are placed back in the sound, and the correction survives transcribing again; Correct Everywhere sits beside it. Each correction takes about two seconds.
• Teach it your vocabulary. Add a term with the ways it gets misheard, and Sharpy fixes it in the captions, the WORDS view and what the agent reads, without hearing the recording again and without touching any other word. With five terms added, every one came out right in all three takes of a test recording. Corrections you make teach the dictionary too.
• Give it the script. If the speaker read from a page, the page corrects spellings, under one rule: it may correct a word and may never supply one.

What can you do with the transcript?

• Edit the video by its words. In the WORDS tab, select words and press Delete, and they leave picture and sound together. Editing a video by editing its text covers it.
• Search what was said. Press ⌘F and type a phrase; the playhead jumps to where it is said, in any recording in the edit.
• Make captions and subtitle files. Tick the captions box in the export panel for SRT and WebVTT beside the video, or burn in animated captions. What is an SRT file explains the formats.
• Translate it. Sharpy translates each caption on your Mac into a subtitle track per language, labelled as machine translation.
• Hand it to an agent. From Claude or Codex, get_transcript returns the transcript as sentences, or word by word on request, to save, summarise or quote.

Sources

Checked on 18 October 2026.

• Apple: Record audio and see a transcript in Notes on Mac
• Apple: View a Voice Memos transcription on Mac
• Apple: Use Live Captions on Mac
• Apple: QuickTime Player User Guide