Extract Subtitles from MKV, WebM & MP4
Pull embedded subtitle tracks out of MKV, WebM and MP4 files as SRT, VTT or ASS — read directly in your browser, so even a 15 GB file needs no upload.
Most tools that claim to do this make you upload the video — a non-starter for a multi-gigabyte remux. Subtitle Doctor reads the container structure directly from your disk: it walks the file index, skips past the video and audio, and reads only the subtitle bytes. That is why a 15 GB MKV finishes in seconds with nothing transmitted anywhere. Every embedded text track is listed with its language and codec, previewable against the video itself, and downloadable as SRT, VTT, or the original ASS with styling intact.
Worth knowing
- This extracts embedded subtitle tracks (separate streams inside the container). Subtitles burned into the picture itself are pixels, not data — recovering those requires OCR of the video, which is a different job.
- MKV text tracks (SRT, ASS/SSA, WebVTT codecs) extract losslessly; ASS tracks are rebuilt as complete .ass files including their style header from the container.
- DVD and Blu-ray sourced tracks (VobSub, PGS) are image-based: the tool detects and lists them, and in-browser OCR support for converting them to text is planned.
- MP4 subtitle tracks use the mov_text (tx3g) or WebVTT format — both are supported; TTML (stpp) tracks are detected but not yet decoded.
- The video codec does not matter for extraction: an H.265 MKV your browser cannot play still extracts fine, because only the container is read, never the video stream.
FAQ
How can this handle a 15 GB file without uploading it?
The file never goes anywhere. Your browser can read local files in small slices, and subtitle data is a tiny fraction of a video file — the tool reads the container headers plus just those subtitle bytes, skipping the gigabytes of video and audio entirely.
My MKV has subtitles but they show as "image-based" — what does that mean?
Tracks that came from DVDs (VobSub) or Blu-rays (PGS) store pictures of the text rather than the text itself. Turning them into SRT requires optical character recognition, which is on the roadmap to run in-browser. Text-based tracks in the same file extract normally.
The tool found no subtitle tracks, but I can see subtitles when I play the video.
Then the subtitles are either burned into the picture (hardcoded — no tool can extract those as text without OCR of the video) or your player is loading an external subtitle file from the same folder. Check for a matching .srt next to the video.
Does this work for WebM files too?
Yes. WebM is Matroska under the hood, and its WebVTT subtitle tracks (both the older D_WEBVTT and newer S_TEXT/WEBVTT conventions) are supported.