How to transcribe video to text on Mac: MP4, MOV to transcript

Transcript or subtitles? Decide first

A transcript is readable text: paragraphs, usually with a timestamp every so often, meant to be searched, quoted and summarised. Subtitles are short timed cues (an SRT file is the common format) meant to be shown over the video by a player. The same speech can produce both, but they're laid out differently, and most tools export one or the other.

• You want a transcript if: you're writing notes, pulling quotes, summarising a recorded meeting or lecture, or making the content searchable.
• You want subtitles if: you're publishing the video and need captions in a player, or an SRT to upload alongside it.

Otter's video page lists TXT, DOCX, PDF and SRT exports. QuietNote produces a Markdown transcript with timestamps; it isn't a subtitle editor.

What macOS offers for video

Apple's own transcription lives in Voice Memos and Notes, and both transcribe audio they recorded themselves, not a video file you open. The one macOS feature that applies to any video is Live Captions.

• Live Captions: Apple's guide says that by default it transcribes audio from all apps, so playing a video in QuickTime or a browser puts captions in a floating window. It needs a Mac with Apple silicon, audio is processed on device, and it's available in a listed set of languages (English, French, German, Japanese, Korean, Spanish and Chinese variants, by region). It's a live view: Apple's page doesn't describe saving the captions, and it says accuracy may vary. Good for following along; not a way to get a file.
• Voice Memos and Notes: transcribe their own recordings on supported Macs (Apple silicon for Voice Memos on macOS Sequoia; M1 or later for Notes). Neither imports a video.

Uploading the video to a hosted service

Otter's video-to-text page says you can upload a file (MP4, MOV, M4A and other common formats) or play the video aloud while Otter listens through the microphone. Its free plan is described as 300 minutes of video transcription a month, and the pricing page lists 3 pre-recorded file imports per user, lifetime, on the free plan; Pro ($16.99 per user per month, or $8.33 billed yearly) allows 10 imports a month and 1,200 imported minutes. Speech recognition is offered in English, Spanish and French. The page says most videos are transcribed at about playback speed: a 10-minute video takes around 10 minutes.

The cost is the upload. A video is often the most sensitive kind of recording you have: it shows faces, screens and whatever was on the desktop during a screen share. If that matters, keep the file on the Mac.

Transcribing the video on the Mac

QuietNote is a note taker for Mac. Import a common video or audio file and it's transcribed on your Mac with the Whisper model that ships with the app, offline. The output is a Markdown transcript with timestamps, saved to a folder you choose. No account, nothing uploaded. If you want a summary or action items, that's optional: a local model through Ollama, or Claude with your own API key, which receives only the transcript text and only when you ask.

1. Import the MP4, MOV or other common video file.
2. Wait for the transcript. A long video takes longer; how long depends on the Mac and the model size.
3. Open the Markdown file, then summarise or pull quotes as needed.

The first signed Mac release isn't out yet; leave your email to hear when it is. QuietNote needs macOS 14.2 or later.

Videos that transcribe well, and ones that don't

• Screen recordings and recorded calls: usually fine. The audio is a direct feed, so there's little room noise. If several people spoke on one track, expect to label the speakers yourself; a single mixed track doesn't carry who said what.
• Phone videos: the microphone is far from the speaker and picks up the room. Accuracy drops with distance and noise, whatever tool you use.
• Music and sound effects under speech: a model may transcribe the words or lose some. Clean dialogue is best.
• Stereo soundtracks: recognition runs on a mono mix. Only if the left and right channels carry different speakers is it worth splitting them first.
• Mixed languages: Whisper is multilingual, but a video that switches language mid-sentence is harder than one language throughout. Whisper's README notes its English-only model variants tend to do better for English.

No tool has a trustworthy accuracy number for your particular video. Apple says of its own Live Captions that accuracy may vary; treat every transcript as a draft to be read once.

What to do with the transcript

• Summary and action items: if the video is a recorded meeting, the meeting summary template gives the transcript a shape.
• Quotes with timestamps: note the time code next to each quote so you can jump back to the frame.
• Chapters: scan the timestamps for topic changes and turn them into a chapter list for the video's description.
• Subtitles: if you need an SRT, take the transcript to a subtitle editor or use a service that exports SRT directly.