How to Transcribe a Video in 2026: Mac, Windows and Online

To transcribe a video, give the file to a tool that pulls out the speech and turns it into text. On a Mac, MacWhisper does it on your computer for free. In a browser, Clipchamp's free autocaptions and Otter's free plan both accept video files. Microsoft 365 subscribers can upload an MP4 to Word's Transcribe pane.

The video is usually sitting right there: a screen recording on the desktop, a client's clip in your downloads folder, a webinar replay you saved. What you want is the one sentence someone said in minute twelve, without scrubbing back and forth to find it. Below: the quickest route for each kind of video file, what each free option really allows, how to get a clean text file out of caption formats, and how to fix a transcript that comes back empty.

Which way to transcribe a video?

Pick the method by where the video lives and how private it is. The limits are taken from each maker's help or pricing pages, linked in the sections that follow. On a Windows PC, Word, Clipchamp and Otter all run in the browser, and Whisper runs from the command line.

If you only need the words from a YouTube video, you may not need any of this: YouTube often has a transcript already. Our YouTube transcript walkthrough covers that. For calls happening right now, our guide to live meeting transcripts is the better fit.

How to transcribe a video on a Mac

MacWhisper transcribes video files on a Mac without uploading them. Its website lists MP4 and MOV among the supported formats, alongside audio such as MP3 and M4A, and says it uses local models to transcribe your files. There's a free download, and a Pro license costs €64 once, with lifetime updates.

1. Download MacWhisper from macwhisper.com and open it.
2. Open your video file in MacWhisper.
3. Pick a transcription model and the spoken language if you know it.
4. Start the transcription and wait for the text to fill in.
5. Export the text, or export subtitles as .srt or .vtt if the video needs captions.

Local transcription is what I'd reach for with anything confidential. When a client sent me a video of their leadership team talking through a reorganization, uploading it to a web service wasn't an option I wanted to explain later. Keeping it on my own laptop meant the question never came up.

One thing to know: MacWhisper's site also mentions connecting to cloud AI services. The privacy benefit holds only while you use a local model, so check which one is selected before you start.

Free and open source: Whisper in the terminal

If you're comfortable with a command line, Whisper, the open-source speech recognition model, is free to run yourself. The Whisper repository says the code and model weights are released under the MIT License and that you need the command-line tool ffmpeg installed. Whisper reads every file through ffmpeg (you can see it in the audio loader), which is why a video file works as input without converting it first:

whisper screen-recording.mov --model turbo

Add --language English (or the language spoken) if the opening is music or silence: when you don't name one, Whisper detects the language from the first 30 seconds. The repository also notes that the turbo model isn't trained for translation, so pick a multilingual model such as medium if you want the speech translated into English.

How to transcribe a video in Microsoft Word

Word's Transcribe feature accepts MP4 files, but not MOV. Microsoft's Transcribe help page lists .wav, .mp4, .m4a and .mp3, allows 300 minutes of uploads a month with a Microsoft 365 subscription, and says it works only in the new Microsoft Edge and Chrome.

1. Sign in to Microsoft 365 in Edge or Chrome and open a document in Word.
2. Go to Home > Dictate (the drop-down arrow) > Transcribe.
3. In the Transcribe pane, select Upload audio.
4. Choose your MP4 or M4A file and wait for the transcript.

Turn a MOV into an M4A first

iPhone videos and Mac screen recordings are usually MOV files, which Word doesn't list. On a Mac, QuickTime Player strips the audio out in a few clicks, and Apple's QuickTime guide says the result is an MPEG-4 audio file with an AAC track, which is an M4A.

1. Open the video in QuickTime Player.
2. Choose File > Export As > Audio Only.
3. Name the file, choose where to save it, and click Save.

A client once sent me a video testimonial straight from her iPhone, and Word wouldn't take it. Exporting the audio in QuickTime took less time than finding out why, and the M4A was a fraction of the size, so the upload was quicker too. That trick helps with any upload tool, not only Word.

How to transcribe a video online with Otter

Otter accepts video uploads, including MP4 and MOV, according to its video to text page, and adds speaker labels and timestamps. For files, though, Otter's free tier runs out fast. Otter's pricing page lists 300 minutes a month, 30 minutes per conversation and 3 lifetime imports per user, with export as .txt.

That makes the free plan a trial more than a workflow: three short clips, then you pay. Otter Pro costs $8.33 a month for each user on yearly billing, or $16.99 paid monthly, and raises that to 10 imports a month and 90 minutes per conversation, per the same pricing page. For a one-off 45-minute video, the free plan won't cover the whole thing.

How to transcribe a video in Clipchamp

Clipchamp, Microsoft's video editor, writes a transcript of any video you put on its timeline, and its pricing page lists AI subtitles on the free plan. The personal version runs in a browser at app.clipchamp.com. The steps below follow Microsoft's autocaptions guide for personal accounts:

1. Import your video from My media on the toolbar and drag it onto the timeline.
2. Click the captions icon in the video preview, or the captions tab in the property panel.
3. Choose the language spoken, then click Transcribe media.
4. Fix any wrong words in the transcript in the captions tab.
5. Click Download captions to save the transcript as an .srt file.

Two limits from the same guide: one transcript language per video, and the download is an SRT caption file rather than a document. Work and school accounts find the download under the … menu in the transcript tab, as Download as .srt.

How to turn an SRT file into plain text

An SRT file is the transcript cut into numbered caption chunks, each with a start and end time. It opens in any text editor (TextEdit on a Mac, Notepad on Windows). For a readable transcript, delete the number lines and the timestamp lines and join what's left into paragraphs. Each chunk looks like this:

12
00:01:04,200 --> 00:01:07,900
So the plan is to launch in March,
not February.

Keep the SRT too. If you ever need to find where a line was said, the timestamp is right there.

How to transcribe a screen recording

A screen recording is a normal video file, so any method above works, as long as the recording has sound. On a Mac, Shift-Command-5 opens the Screenshot toolbar, and Apple's screen recording guide says recordings are saved as .mov files on the desktop by default. Sound is only captured if you choose a microphone under Options; Apple's page says newer macOS versions also offer Include System Audio.

I learned that the hard way. I recorded a walkthrough of a new dashboard for my team, talked through the whole thing, and only noticed afterward that no microphone was selected. The file was silent, so there was nothing for any tool to transcribe. Now I check Options before I press record.

• Loom: transcripts come with the video. Loom's help page says they're on all plans in more than 50 languages; open Transcript in the side panel and use the copy button. Downloading captions as an SRT needs a Business, Business + AI or Enterprise plan.
• Clipchamp: its free plan includes a screen recorder, per its pricing page, so you can record and transcribe in the same place.
• Recorded Zoom, Teams or Google Meet calls: the meeting app may already have made a transcript, as our live transcription guide explains. Check there before you transcribe the file again.

How to transcribe a TikTok, Instagram or YouTube video

For your own social videos, transcribe the original file in your camera roll or Photos library, not a copy saved back from the app. The original has the cleanest sound, and you already have it. When I post a short product clip to Instagram, the version on my phone is the one I transcribe for the caption and the blog post that goes with it.

For someone else's video, start with what the platform already shows: on YouTube that's the transcript panel, explained in our YouTube transcript guide. Downloading the file is a different matter. YouTube's Terms of Service, for example, don't allow downloading content unless the service itself permits it, YouTube or the rights holder gives written permission, or the law allows it. If you need the words for a quote, the transcript panel plus a link to the source is the cleaner route.

Why is my video transcript blank or full of errors?

An empty or garbled transcript almost always comes down to the audio, not the tool. Go through these in order:

1. Play the file with the volume up. If you hear nothing, the video has no audio track (a common screen recording slip), and no tool can fix that.
2. Check the format. Word lists only .wav, .mp4, .m4a and .mp3; export a MOV to M4A in QuickTime first.
3. Set the language yourself. Whisper, for one, guesses the language from the first 30 seconds, so a music intro can send it the wrong way. Where a tool offers a language setting, as Clipchamp and Whisper do, use it.
4. Trim long silences and intros. YouTube's own caption help page names a long silence at the start, poor sound and overlapping speakers among the reasons automatic captions fail, and the same things trip up other tools.
5. Split very long files. Free plans cap length (Otter's at 30 minutes per conversation), so the end of a long video may simply be missing.

Clean up the transcript before you use it

Machine transcripts are a first draft. Before I paste a line into a deck or an email, I play the moment back and check it against the video, because a dropped "not" turns a quote upside down. A few habits make the cleanup quicker:

• Name the speakers where the tool only says Speaker 1 and Speaker 2. Otter adds speaker labels; Whisper's output is words and timestamps without names.
• Check names and numbers first. Product names, people's names and figures are the details people quote, so a wrong one does the most damage.
• Keep the timestamps in your working copy, even if the final version drops them.
• Tell people before you record. If the video is a conversation you're recording yourself, the consent rules in our free transcription software guide apply here too.