Subtitle transcript

SRT to plain text (TXT)

Extract text from SRT or VTT subtitles without timecodes: strip cue numbers, style tags, and YouTube auto-caption repeats, then copy or download UTF-8 TXT.

Cleans format only—no transcription, translation, summary, or rewriting

Subtitle track

Extract an editable transcript from the timeline

SRT · 4 cues

Drop or choose SRT / VTTUTF-8 text · 5 MB max · or paste into the box belowSubtitle content

Text cleanupRemove style and speaker tagsRemove “Name:” prefixesMerge consecutive duplicate linesLeave blank lines between cues

Sample SRT loaded. Convert it now or paste your own subtitles.

Result proof

224-character transcript

UTF-8

Transcript preview

Some tools give you one button.

We would rather make the timing, format, and limits clear.

LOCAL FIRST: subtitle text never leaves your browser.

Speaker: Sync can shift a whole track or correct drift that grows over time.

Output stays UTF-8. WebVTT uses a period for milliseconds; SRT uses a comma.

Methods & limits

A transcript is not made by deleting every non-letter character

Line breaks inside captions may exist only for screen width; timecodes and cue numbers are not prose. The tool joins lines within each cue, optionally removes tags and speaker prefixes, and preserves paragraph rhythm.

Deduplication compares neighbouring cues only: an identical cue is dropped, and when a cue starts with the previous cue's last line—the rolling style of YouTube auto-captions—only the new line is kept. Matching sentences elsewhere are never removed, and no semantic guessing is done.

Parse cues → clean tags → optionally drop neighbouring repeats and rolling overlap → join within-cue lines → optionally remove speaker prefixes → output UTF-8 TXT.

Typical tasks

SRT to plain text (TXT) · Typical tasks

Each example lists the job, the result you get, and what still needs a human check against your own files.

Turn a lecture SRT into notes

Task: An online course ships an SRT file and you want the words without numbers or timestamps.
Result: Each cue becomes one line or paragraph with in-cue line breaks joined, ready to paste into a document.
Check: Breaks follow cue boundaries, not sentences; punctuation comes only from the original captions.

Clean up YouTube auto-captions

Task: The downloaded auto-caption .vtt shows every line twice and carries word-timing tags.
Result: Word-timing tags are removed and rolling repeats collapse, so each spoken line appears once.
Check: Auto-captions usually lack punctuation and speaker names; the tool does not add them.

Quote dialogue without speaker names

Task: Captions prefix each line with a name such as “JOHN:”.
Result: Turn on Remove “Name:” prefixes to keep only the spoken words.
Check: Any opening of up to 24 letters or digits before a colon is removed, including text like “Note:”. Read before publishing.

FAQ

SRT to plain text (TXT) · FAQ

How do I convert SRT to TXT without timecodes?

Drop or paste the .srt file. Timecodes and cue numbers never enter the output; Download TXT saves a UTF-8 text file.

Does it work with YouTube auto-generated captions?

Yes, with the .vtt file: word-timing tags are removed, and with “Merge consecutive duplicate lines” on, each rolling line is kept once. It does not fetch captions from YouTube; you need the .vtt file first.

Can it transcribe a video directly?

No. It processes existing SRT or VTT text, uploads no audio or video, and performs no speech recognition.

Why remove only neighbouring repeats?

The same sentence may matter in different parts of a document. Handling only exact repeats and rolling overlap between adjacent cues reduces accidental deletion.

Are line breaks preserved?

Lines inside one cue become spaces. Separate cues can appear one per line or with blank lines between them.