Verbatimly

Notes on transcription

Updated 15 November 2026

Two things decide whether a transcript is usable: the format you save it in, and the recording you started from. Both are easy to get right before the fact and annoying to fix afterwards.

SRT vs VTT vs TXT: which subtitle format

SRT is the safe default — every editor and player reads it. VTT is the web standard and supports styling that SRT cannot. Plain text has no timestamps at all, which is the right choice when you want to read or search the transcript rather than watch it.

The differences that actually bite: SRT separates milliseconds with a comma, VTT with a period, and most "my subtitles are off by a frame" problems start there.

Read: which subtitle format to use →

What to fix before you transcribe

Automatic transcription is sensitive to three things: overlapping speakers, music under the voice, and distance from the microphone. None of them can be repaired after the fact, and all of them are cheap to avoid when you press record.

Read: preparing a recording for transcription →

Why subtitle timestamps drift

Subtitles that start in sync and end up seconds late have a specific cause, and it is almost never the player. The four things that produce drift, how to tell them apart from a constant offset, and the order to check them in.

Read: why timestamps drift →

The subtitle format family

Six formats, six different ideas about how to write down "what text appears when". SRT, WebVTT, SBV, SAMI, MicroDVD and ASS — what each one looks like inside, who made it, and which to pick.

Read: a guide to the formats →

Try it

Transcribe a file in the browser → — nothing is uploaded except the compressed audio track.