TL;DR: The fastest way to study a video course is to get its words out as text, then take notes from that instead of scrubbing the video. How you get the text depends on the course: if it has captions, pull the subtitle track; if the subtitles are burned into the picture, OCR them; and if there are no subtitles at all, transcribe the audio with speech recognition. Once it is text you can search it, highlight it, and — if the course is in another language — translate it.
Why turn a course into text first?
Video is a slow way to review anything. You cannot search it, you cannot highlight a spoken sentence, and finding the one point the instructor made ninety minutes in means dragging a playhead back and forth until you hit it. Getting the course's words out as timed text changes the medium entirely: hours of video become a document you can skim in minutes, search by keyword, and quote straight into your notes. Once the words are text, the timeline stops mattering.
The goal is to study the text, not the timeline — turning a course into words you can search and quote is what makes note-taking from video fast.
If the course has captions or a subtitle track
Many course players and course downloads already include a caption track — the same captions you can toggle on during playback. When that track exists, there is no reading or OCR involved: you just pull the text out. Export the transcript from the platform if it offers one, or pull the subtitle track out of the video file as an SRT with ffmpeg. Either way you have the full text of the course in seconds.
If the course already ships with captions, extracting the text is the easy case — pull the subtitle track as an SRT and you have the words instantly, no OCR needed.
If the subtitles are burned into the video
Some course rips have the subtitle text baked directly into the picture — the words are part of the video frames, not a separate track you can toggle. There is nothing to pull out with a caption export or ffmpeg, because there is no text inside the file; it is a burned-in subtitle. To get the words, you OCR the frames: read the on-screen text out of the picture and turn it back into an SRT. GeekLink does this locally on a Mac, detecting the subtitle region and reading each frame into clean, timed lines.
When the words are burned into the picture, no exporter can pull them out — you OCR the frames to read the on-screen text back into an SRT.
If the course has no subtitles at all
The hardest case is a lecture with nothing on screen: no caption track, no burned-in text, just a person talking. There is no subtitle to extract because one was never made. The way to get the words is to transcribe the audio with speech recognition — let the software listen to the lecture and write down what was said. GeekLink uses Whisper on-device to turn the spoken lecture into a timed transcript you can then edit and clean up.
No subtitles anywhere means there is nothing to extract — transcribe the audio with speech recognition to turn the spoken lecture into a timed transcript.
Turning the text into notes (and translating a foreign course)
However you got there — a pulled subtitle track, OCR'd frames, or a transcript — you now have an SRT or a plain transcript, and that is a text document you can work with. Search it for the terms you care about, pull the key lines into your notes, and strip out the timestamps if you just want prose. If the course is in another language — say a French course you are studying in English — translate the subtitles into your language first (GeekLink translates into 40+ languages) so your notes come out in the language you actually think in. All of this runs locally on your Mac, from the OCR and speech recognition to the translation and the SRT export.
Once the course is text, notes are just a matter of searching, quoting, and — for a foreign course — translating into your language before you write anything down.
FAQ
How do I get the text out of a video course?
If it has captions, pull the subtitle track; if the text is burned into the picture, OCR it; if there are no subtitles, transcribe the audio with speech recognition.
The course has no subtitles — can I still get notes?
Yes — use speech recognition to transcribe the audio into a timed transcript, then take notes from that.
The course is in another language. How do I make notes in mine?
Extract or transcribe the subtitles, then translate them into your language before you take notes.
Can I get a plain transcript without timestamps?
Yes — export the SRT and strip the timing, or copy the text lines into your notes document.
Is this legal / does it upload my course?
GeekLink runs locally on your Mac with no upload; what you do with your own study notes is up to you.