Find the Line, and the Timecode It Came From
Paste a transcript or drop an SRT or VTT file, type the word you are looking for, and every match comes back with the timestamp attached. Not “somewhere in the file,” but 12:04, on this line, with your term highlighted in context.
It runs in the browser. The transcript is parsed and searched on your device, which matters when the transcript is an unpublished interview or an internal call.
Why Not Just Use Ctrl and F
Because Ctrl+F finds the word and then leaves you stranded. It will highlight “pricing” in the middle of a wall of text, and you still have no idea where in the recording that was. The timecode is the thing you actually needed.
Subtitle files make it worse. Open an .srt and it is half scaffolding: a sequence number, a timing line with arrows, then two lines of text, over and over. Ctrl+F matches inside that mess and you are counting backwards to find which cue you landed in. This parses the cues first, so the text is clean to read and each match already knows its own start time.
It also joins the wrapped lines back together. Caption files break a sentence across three cues, so a phrase like “usage based pricing” can be split in the file and Ctrl+F simply will not find it. Here the cue text is stitched before searching, so the phrase matches the way you would say it.
Searching a YouTube Transcript
YouTube will show you a transcript and then give you no way to search it. Open a video, hit “Show transcript” under the description, and you get the whole thing in a scrolling panel with timestamps, and no search box anywhere in it. On a forty minute video that panel is close to useless if you are hunting one sentence.
The way around it is three steps. Select the transcript text in that panel, copy it, and paste it here. YouTube’s transcript comes with its timecodes attached, so every match you get back still points at the right minute of the video. Then search the word.
It works the same on auto-generated captions, which is most of YouTube. Those are the ones worth searching, because auto-captions are exactly where you cannot trust your memory of the wording, and they carry the timestamps you need to jump back to the moment.
On a Phone There Is No Ctrl and F
This is the version of the problem nobody has a good answer for. On a laptop you can at least fall back on Ctrl+F. On a phone there is no find-on-page in the YouTube app at all, and the transcript panel is a thumb scroll with no search in it.
So on mobile the flow is the same three steps and it saves more time, not less: open the transcript in the app, copy it, open this page in your phone browser, paste, search. The page works at phone width, and because the search runs on the device you are not pushing a transcript over mobile data to find one word.
Exact Match, on Purpose
This is literal string matching. It finds the characters you typed and nothing else. That is a deliberate limit, not a missing feature.
Exact search is what you want when you know the word: a product name, a number someone said, a person’s surname, a phrase you are quoting. It is repeatable and you can trust the count, which is why journalists and researchers reach for it.
It is the wrong tool when you do not know the word. If the job is “find the part where they talk about churn” and nobody said the word “churn,” exact matching returns nothing while the moment is right there. That is a semantic search job, and it belongs on the video finder, which indexes the video itself and finds moments by meaning rather than by spelling. Two different tools for two different questions, and it is worth knowing which one you have.
Whole Word and Match Case Earn Their Keep
Both toggles look fussy until the first time they save you.
Search “AI” without whole word matching and you get “said,” “again,” “sustainable,” “maintain,” and every other word with those two letters buried inside. Turn whole word on and you get the acronym, only where it was actually spoken. This is the single most common reason a transcript search looks broken.
Match case does the opposite job. Leave it off by default, because transcripts capitalize inconsistently and you do not want to miss a hit because the speaker started a sentence. Turn it on when the capitalization is the point: separating the company “Apple” from the fruit, or an acronym from an ordinary word.
Who Actually Does This
Journalists working from an interview recording, finding the exact sentence for a quote and the timecode so the video editor can pull the clip. The timecode is half the deliverable.
Researchers coding qualitative interviews, counting how many times a term came up across a session, which is a number you can defend because the match is literal.
Support and product teams searching a demo or a webinar recording for where a feature was explained, so they can link a customer to the minute instead of the hour.
And editors hunting a pull quote they half remember, which is where whole word matching and a short phrase beat scrolling every time.
If You Do Not Have a Transcript Yet
This searches text you already have. To create the transcript from a recording first, the transcript generator produces one with speaker labels. If the transcript you have is a mess of timestamps and filler, the transcript cleaner strips that out in the browser. And to turn a finished transcript into a subtitle file, the text to SRT converter handles the timing.
