How to get subtitles from your video

Case 1: the video has an embedded subtitle track

This is the common, easy case. A lot of videos, MKV especially, but also plenty of MP4, MOV, and WebM files, store the subtitles as a separate text track inside the container, sitting next to the picture and sound. Nothing is drawn onto the video itself; the subtitle text is just carried along, which is why a player can switch it on and off or offer more than one language. Because it is already text, you can lift it straight out as a standalone .srt or .vtt file.

The Extract Subtitles tool does exactly this. Drop the video in, it lists the subtitle tracks it finds, and you pick one to save as SRT or VTT. It runs entirely in your browser and reads only the subtitle track, so the video is never uploaded and even a large file stays on your device.

Case 2: the subtitles are burned into the picture

Sometimes the subtitles are part of the video image itself, painted onto every frame, so they are always visible and can't be turned off. These are called hardcoded or burned-in subtitles, and there is no text to extract, because there is no separate track. The words only exist as pixels.

Turning burned-in subtitles back into a text file needs OCR (optical character recognition), the same technology that reads text out of a scanned document, run frame by frame across the video. That is a heavier, separate process with its own accuracy trade-offs, and this site doesn't do it. It is worth knowing so you don't spend time looking for a track that isn't there.

Case 3: the video has an image-based subtitle track

There is a middle case. Some files, often those made from discs, carry the subtitles as a real separate track you can toggle, but the track stores each line as a little picture rather than as text. PGS and VobSub are the usual formats. A player can show and hide them like any track, which makes them look like Case 1, but because each subtitle is an image, they can't be saved as text directly.

Like burned-in subtitles, image-based tracks need OCR to become an SRT, so they fall outside what a text extractor can do. The Extract Subtitles tool recognises these and tells you the track is image-based rather than producing an empty or broken file, so you are not left guessing.

How do I tell which one I have?

Play the video and look for a subtitle or caption control. If you can switch the subtitles off, or choose between languages, they are a separate track, which is Case 1 or Case 3. If they can be turned off and you can extract them to text, it was Case 1; if the extractor reports them as image-based, it was Case 3. If the subtitles are always on screen and there is no way to turn them off, they are burned into the picture, which is Case 2 and needs OCR. This one check tells you in a few seconds whether there is anything to extract.

Once you have the subtitle file

After you have pulled out an SRT or VTT, the other tools here pick up where extraction leaves off. If the subtitles run ahead of or behind the audio, line them back up with the Shift tool. If you need a different format, convert with SRT to VTT or VTT to SRT. And if the text is cluttered with sound descriptions and speaker labels, strip them with Remove SDH, or tidy auto-generated YouTube captions with Clean YouTube Captions.