A timed transcript is useful while checking a recording, but the same timestamps can interrupt a reading copy or an article excerpt. Remove only the timing prefixes you have identified, while keeping speaker names, paragraphs and any times mentioned in the conversation.
This procedure uses Find and Replace Text on a plain-text transcript with one MM:SS timestamp at the beginning of each speech line. Keep the original timed version so you can return to the recording later.
Check the timestamp format first
The example expects two minute digits, a colon, two second digits and at least one space or tab before the speaker. The minute field can run from 00 to 99; the seconds must be 00 through 59. These are elapsed-time prefixes, rather than times spoken by a participant.
Inspect several lines before choosing the pattern. Bracketed timestamps, three-digit minutes, hours and timing lines from subtitle files need a different procedure. This edit does not convert SubRip or WebVTT files, and it does not check the transcript against its audio.
Try a transcript with a time inside the dialogue
Paste this fictional sample into Your text, with a blank line before the final speaker turn and no extra newline at the end:
00:03 Mira: Welcome to the workshop.
00:18 Jon: The oven timer reads 12:30.
01:05 Mira: Let the bowl cool.There are three timing prefixes to remove. The 12:30 inside the second sentence is part of the dialogue and must stay. A broad search for every digit-colon-digit sequence could remove that information too.
Replace the prefixes and keep the line breaks
- Enable Use regular expression and Replace all.
- Leave Whole word only off. Turn Match case on for a consistent test; it does not change this pattern because the pattern contains no letters.
- Enter the following in Find, with single backslashes and no surrounding slashes or flags:
(^|\r?\n)[0-9]{2}:[0-5][0-9][ \t]+In Replace with, enter:
$1Select Replace. Expect 3 Replacements made, 107 Characters before and 89 Characters after. The result is:
Mira: Welcome to the workshop.
Jon: The oven timer reads 12:30.
Mira: Let the bowl cool.The three prefixes have gone. Both speaker names, the spoken 12:30 and the blank line remain. The original stays in the input box; the edited version appears in the separate result area.
Understand the boundary and replacement
The opening (^|\r?\n) matches either the start of the input or a line break. Its parentheses capture that boundary as group 1. For the first line, the captured text is empty; on later matching lines, it is the line break. Replacing the match with $1 puts that captured text back while removing the timestamp and its following separator.
MDN documents numbered captured-text replacements in its String.replace reference. The site tool supports the $1 token used in this procedure.
The current tool has no multiline-flag control. A pattern beginning with only ^ would target just the start of the entire input here. The explicit line-break alternative lets this pattern reach later lines. See MDN’s input-boundary explanation for the difference between ordinary and multiline matching.
The final [ \t]+ consumes spaces or tabs after the prefix. It deliberately avoids \s+, which also includes line terminators, as described in MDN’s whitespace character-class reference. The narrow separator helps preserve an existing paragraph break.
Investigate unchanged timestamps
This pattern leaves 00:03:05 Mira:, [00:03] Mira:, 100:03 Mira: and an indented 00:03 Mira: unchanged. It also leaves 00:60 Mira: unchanged because the seconds are outside the stated range. Do not loosen the pattern until you understand which format your source actually uses.
Compare the replacement count with your expected timing prefixes, then check the first and last speaker turns and a sentence containing an ordinary clock time. Use the draft comparison guide to review a longer edit. Copy the reviewed reading version into its destination and confirm its paragraph spacing there; retain the timed source for reference.