fixsubtitles

Extract text from subtitles

Strips cue numbers, timecodes and tags to give you clean plain text — optionally without speaker labels and flowed into paragraphs. In your browser, no upload.

Runs in your browser. Files never leave your device.

Drop a subtitle file

or ·

SRT, VTT or ASS · processed in your browser · nothing is uploaded

From timed cues to plain prose

A subtitle file is mostly scaffolding — numbers, timecodes, arrows, tags — wrapped around the words. This tool keeps the words and discards the rest, joining each cue’s internal line breaks into a single line and stripping any <i>/<b> formatting. 

input.srt                         output.txt
1
00:00:01,000 --> 00:00:03,000        Who's there?
Who's there?                          Nobody. Go back to sleep.
2
00:00:03,200 --> 00:00:05,000
<i>Nobody.</i> Go back to sleep.

Speaker labels and paragraphs

Two switches shape the output. Remove speaker labels drops leading names (JOHN:) and the “- ” dashes used for alternating speakers — useful when you want only the words. Join into paragraphs flows consecutive cues together and starts a new paragraph after a clear gap in the dialogue, turning a list of fragments into something you can actually read.

What it’s good for

Pulling a transcript out of a video you’ve subtitled, getting a block of text to translate or summarise, quoting dialogue, or feeding clean text into another tool. It’s the fastest way from a subtitle file to readable copy — with no timestamps in the way.

Limits

The output is only as good as the subtitles: this extracts existing text, it doesn’t transcribe audio. Paragraph grouping uses pauses between cues as a heuristic, so very dense dialogue may stay as one block. Everything runs locally — no upload, instant.

Frequently asked questions

Does it remove the timestamps and cue numbers?

Yes — that’s the whole job. The cue numbers, the timecodes and the formatting tags all come out, leaving just the spoken text as clean plain text you can paste anywhere.

What does “remove speaker labels” do?

It strips leading speaker names like JOHN: and the dialogue dashes (“- ”) that scripts use, so you’re left with just what was said. Leave it off if you want to keep who said what.

What’s the “join into paragraphs” option?

By default each cue becomes one line. With paragraphs on, consecutive cues are flowed together and a new paragraph begins after a noticeable pause in the dialogue — giving you readable prose instead of a list of fragments. Good for turning subtitles into a transcript.

Can I use this to get a transcript for notes or translation?

Yes. Extracting the text is a common first step for summarising, translating, or quoting a video. If the video has no subtitles to begin with, you’ll need them transcribed first.

Is anything uploaded?

No. The extraction happens in your browser; your file never leaves your device.