SubToolbox

SRT to TXT — Turn Subtitles Into Clean Text

Strip timings and numbers from SRT or VTT to get clean plain text — optional paragraph merging for readable transcripts.

Quick answer: Drop in an .srt or .vtt and get the spoken text with all timings and numbers removed. “Group into paragraphs” merges cues separated by less than 2 seconds into readable paragraphs — ideal for transcripts and articles.

Drop a subtitle file here, or browse — it stays on your device

Core facts
InputSRT or WebVTT (auto-detected)
OutputPlain text — one cue per line, or merged paragraphs
StrippedNumbers, timestamps, HTML tags, {\tags}, (sound) labels optional
Paragraph ruleNew paragraph after any pause ≥ 2 seconds
PrivacyRuns in the browser; no upload

What this tool does

Subtitles make poor reading material: numbers, arrows and timestamps interrupt every sentence. This converter extracts just the words — cue numbers and timings dropped, stray HTML or ASS tags stripped, optional bracketed sound labels like [ Applause ] removed. With paragraph grouping on, cues separated by short gaps flow into one another and a pause of two seconds or more starts a new paragraph, which turns a caption file into something that reads like notes from the talk. The result is copyable and downloadable as .txt.

Common uses

  • Turning a downloaded caption file into a readable transcript
  • Extracting quotes or content from a talk's subtitles for an article
  • Preparing subtitle text for translation in a plain-text workflow
  • Feeding lecture captions into notes or a summary tool

Frequently Asked Questions

Will duplicate lines be removed?
Consecutive identical cue texts (common in karaoke-style or rolling captions) are kept by default so you can verify the extraction; use the merge option to smooth them into flowing paragraphs.
Does it work with VTT files?
Yes — the format is auto-detected, so .srt and .vtt (and SBV/ASS too) all convert to text the same way.
Can I keep the timestamps in the output?
Not on this page — its job is removing them. If you want a timestamped transcript, the SRT file itself is already the closest form; export each cue as “00:01:23 text” with a quick find-and-replace after converting.

More subtitle converters