Convert SRT & VTT Subtitles to Markdown

What it cleans up

Both formats carry the same noise around the text. The converter removes it:

• Timestamps and cue numbers are dropped (or kept as a compact inline prefix, your choice).
• Rolling / scrolling captions (common in auto-generated YouTube subtitles) repeat each line across consecutive cues. The converter dedupes that so each sentence appears once.
• Inline styling is translated: <i> becomes *italic*, <b> becomes **bold**; karaoke and word-timing tags are removed.
• Cue text is reflowed into paragraphs at natural pauses, so the result reads like prose instead of caption fragments.

Options options

• Keep timestamps (off by default): prefixes each block with an inline [hh:mm:ss] marker, useful for citing or navigating a long transcript.
• Preserve speaker labels (on by default): WebVTT voice tags like <v Alice> become **Alice:** turns, and a new speaker starts a new paragraph.

Limits & privacy

• Auto-caption cleanup is heuristic. The dedupe handles the common rolling-window pattern; an unusual caption layout may leave a stray repeat to tidy by hand, and a phrase deliberately repeated within a few seconds can occasionally be merged.
• Sentence boundaries are detected on punctuation and pauses, so an abbreviation like "Mr." can occasionally split a paragraph early.
• Privacy: the file is parsed entirely in your browser. Nothing is uploaded, and there is no size cap beyond what your browser can hold (very large files are bounded for responsiveness).