Why recording software prefers MKV
MKV is the container of choice for people who record for a living, and the reason is unglamorous but important: it survives a crash. Because of how it stores data as it goes, a recording interrupted by a power cut or a frozen machine usually remains playable, where the same interruption can leave an MP4 unreadable.
That reliability is why OBS, the tool behind most streams and a large share of screen recordings, writes MKV by default and suggests converting to MP4 only after the recording is safely finished. If you record anything long or important, there is a good chance MKV is what you have.
Multiple audio tracks, and which one we read
The other thing MKV does well is carry many tracks in one file. A streamer's recording might hold the microphone on one track, desktop audio on another and a guest's voice on a third, all bundled together so they can be balanced later in an editor.
We transcribe the first audio track in the file. When your recording separates speakers that way, exporting the track you actually want as its own file gives a noticeably cleaner transcript than a mixed version, because the engine is no longer hearing game audio or music underneath the speech.
Frequently asked questions
Do I need to convert MKV to MP4 before uploading?
No. Remuxing to MP4 is a step people take for compatibility with editors and players, and it does nothing for transcription. Upload the MKV directly and skip it.
My MKV has separate microphone and desktop tracks. Which is used?
We read the first track. For the cleanest result, export just the microphone track from your editor and upload that, so background game or music audio is not competing with the speech.
Will subtitles already inside the MKV be used?
No. If your file already carries a subtitle track, that text came from somewhere else and we ignore it. We transcribe the audio itself and give you a fresh transcript with its own timestamps.