Turn SRT and VTT into plain text

What gets removed

A subtitle file is mostly structure around a small amount of text. Take this one.

1
00:00:01,000 --> 00:00:03,500
<i>Welcome</i> to the show.

2
00:00:04,000 --> 00:00:06,000
JOHN: Thanks for having me.

With the default settings it becomes:

Welcome to the show.
JOHN: Thanks for having me.

Tick Remove speaker markers and the second line loses its JOHN: label. The speaker option is deliberately narrow: it removes >> turn markers, a leading hyphen, ALL-CAPS labels followed by a colon, and VTT <v Name> tags. It does not touch “Note: this matters”, because a mixed-case word and a colon is how ordinary sentences are written.

Choosing a layout

• One line per subtitle joins the two lines of a cue into one and keeps each cue on its own line. Good for searching and diffing.
• Keep the original line breaks preserves the way each cue was wrapped.
• Merge into paragraphs runs the cues together and starts a new paragraph after a pause you choose. This is the one to use if you want something that reads like a document.
• One single paragraph removes every break, handy before pasting into a field that does not accept line breaks.

Clean-up options and when they matter

Formatting tags. SRT files from different tools carry <i>, <b>, <font color> and the positioning codes written like {\an8}. VTT adds <v>, <c> and in-cue timestamps. Leaving the tag option on removes all of them; turn it off if you want to keep italics for an editor that understands them.

Sound descriptions. Closed captions include things like [Music] and (laughs) for viewers who cannot hear them. They are useful in the captions and noise in a transcript, so there is an option to remove anything in square brackets or parentheses. It cannot tell a sound cue from a real aside in parentheses, so read the result if the text uses brackets for anything else.

Repeated lines. Auto-generated captions are the main source of these. Dropping exact consecutive repeats is safe; it will not remove a line you genuinely said twice with something between.

Limits worth knowing

The output is only as good as the captions. Punctuation and capitalisation in a subtitle file are often incomplete, and names or numbers that were wrong in the captions are still wrong here. There is no timing in plain text, so keep the SRT if you will need it later. The same job is covered in more detail in how to convert SRT to plain text, and transcript file formats explained explains what else a transcript file can hold. To count the words in the result use the transcript word counter. If you have the recording but no captions at all, transcribe the video to text directly.