Lemora › Guides

Turn an MKV lecture recording into a transcript on Mac

If you have an MKV lecture recording and need the spoken words for study, start by extracting its audio. Lemora saves an M4A file, which you can then transcribe into plain text. The useful result is a reviewed transcript beside the recording, with the original video still available for diagrams and demonstrations.

Before you start: Transcription requires macOS 26 or later, an available on-device speech engine, and support for the Mac's current language. macOS may need an internet connection once to install its speech model; the recording is not uploaded.

Extract the sound, then transcribe the M4A

  1. Save the MKV in Finder. Listen to enough of the source to confirm that the lecturer is audible and that the recording contains the material you need.
  2. With Lemora running, select the MKV and press Shift-Command-E. On the Tools page, choose M4A (Extract Audio).
  3. Play the new M4A. Confirm that it contains the lecture audio, rather than a commentary or another language track.
  4. Select the M4A in Finder and open the wheel again. On the Tools page, choose Transcribe. Allow macOS to install its speech model if needed.
  5. Open the saved TXT file and compare passages with the audio. Correct technical terms, names, numbers and any wording you intend to quote.

Results are saved beside the originals when folder access is available, or in Downloads/Lemora otherwise. Source files are never overwritten.

Where to find these actions

What you get

An M4A recording and a plain-text transcript

Extract Audio writes the soundtrack as M4A. Transcribe writes a separate TXT file. This workflow does not generate slide notes, a summary or speaker labels.

Works with: An MKV lecture file containing audible speech. Audio extraction also supports MP4, MOV and WebM.

Check speech support before you begin

Audio extraction and speech recognition are separate capabilities. Lemora can extract the sound without making transcription available on every Mac. Transcription needs macOS 26 or later and Apple's on-device speech engine. The current workflow uses the Mac's language; it does not offer a language picker for choosing a different lecture language.

Use the M4A when you only need to listen

For revision during a commute or a task that only needs the soundtrack, the extracted audio may be enough. Transcription is a separate step, so you can decide whether written text adds value. Keep the MKV as well: the lecturer may refer to something shown on a slide that cannot be understood from the speech alone.

Review the terms that matter to the subject

A recognizable sentence can still contain a wrong name or technical term. Listen around definitions, equations spoken aloud, units and quoted figures. Mark unclear passages for another listen instead of filling the gap from memory. Use the transcript to navigate the material, then return to the source before quoting it or relying on an exact statement.

The transcript contains speech, not the lecture slides

Text recognition from slide images is not part of this workflow. A lecturer may say 'as you can see here' without describing the diagram. Add your own notes while viewing the relevant part of the video. The TXT output is plain text rather than a set of synchronized captions or a finished study guide.

When captions are the real goal

If you want text displayed at the right times during video playback, use Make Subtitles on the original video's Tools page. That writes an SRT file. A transcript is easier to read as prose; a subtitle file is organized around time. Choose the output for the way you plan to use it.

Questions

Will Lemora identify different speakers?

This workflow does not add speaker labels. Review changes of speaker yourself if the recording includes questions or discussion.

Does offline transcription mean no download is ever needed?

No. macOS may need to download the supported language model first. After installation, recognition runs on the Mac; your recording is not sent to a transcription server.

Why is Transcribe unavailable?

Check the macOS version and whether the on-device speech engine supports the Mac's current language. Installing Lemora alone does not make speech recognition available on an unsupported setup.

Try this workflow on your Mac

Download Lemora on the Mac App Store Free to try: every feature for your first 25 actions. macOS 15 or later.

More guides