The subtitle-to-text converter pulls just the words out of an SRT or VTT file and gives you a clean, readable transcript. It throws away the sequence numbers, timecodes and markup, and lets you collapse the result into single lines or a flowing paragraph — handy for reading, archiving, or pasting subtitled-video dialogue into a document.
How it works
The file is split into cues. For each block the tool skips the WebVTT header and any NOTE or STYLE
sections, skips the optional numeric index or cue identifier, skips the start --\> end timing
line, and keeps the caption text that follows. That text is then processed according to your
options:
- Strip formatting removes HTML-style tags and ASS-style override blocks.
- Collapse to one line joins a cue’s wrapped lines into a single space-separated line.
- Paragraph mode concatenates every cue into one flowing block of prose.
- Remove duplicates drops a cue identical to the previous one, which cleans up rolling captions that repeat lines as they scroll.
A live word count shows how much dialogue the transcript contains.
When each mode is most useful
Line-per-cue mode (default) keeps a blank line between each caption block. This is the most readable format when the source cues already correspond to complete sentences or natural speech pauses, and it makes it easy to spot where the speaker changed or where a new scene started. Use it for archiving or review.
Paragraph mode joins everything into a continuous block of prose. This is what you want when pasting the transcript into a document, a word processor, or an AI summarisation pipeline. The continuous text flows naturally without the gaps that line-per-cue leaves.
Collapse to one line is most useful for word-wrapped captions where a single spoken phrase was split across two physical lines in the SRT file. Without collapsing, each half becomes its own line in the output.
Remove duplicates is essential for rolling or “karaoke” captions, where the same phrase appears across several consecutive cues as the text scrolls. Without de-duplication these transcripts are almost unreadable, with each phrase repeated four or five times in a row.
Practical uses for the extracted transcript
- AI summarisation — paste the transcript into an LLM to get a chapter summary, key takeaways, or Q&A pairs from a lecture or documentary without watching the full video.
- Search and citation — a plain-text transcript is searchable with standard tools, so you can find and quote specific passages from hours of video content.
- Accessibility archiving — storing a transcript alongside a video ensures the content remains accessible even if the caption file format becomes unsupported later.
- Translation input — a clean transcript is easier to feed into a translation tool than a raw SRT file, and the result can be re-imported as a new caption track.
Everything runs locally in your browser — no upload, no account required.