Subtitle to Transcript Converter

How to use

1. 1Click Open .srt or .vtt to load a subtitle file from your device, or paste the contents directly into the input on the left. Use Load SRT sample or Load VTT sample to see a worked example.
2. 2Pick a paragraph mode: Merged paragraphs groups consecutive cues into prose, Line per cue keeps the original cue boundaries, and Single block produces one continuous paragraph. In Merged mode, slide the pause threshold to control where paragraphs break.
3. 3Choose a timestamp mode: None for clean prose, Per paragraph for navigable section markers, or Per cue for a fully timestamped reading copy.
4. 4Pick a speaker-label mode for WebVTT voice tags: Strip removes them, Inline keeps the name as a prefix in the same line, and Prefix line puts the name on its own line above each paragraph.
5. 5Toggle Strip styling tags, Deduplicate repeats, and Smart sentence case to clean up auto-captions and produce readable prose.
6. 6Switch the output format to Plain text or Markdown, then click Copy or Download to grab the final transcript as a .txt or .md file.

About this tool

Subtitle to Transcript Converter parses any SubRip (.srt) or WebVTT (.vtt) file and produces a clean plain-text or Markdown transcript ready for a blog post, course handout, translation pass, podcast description, accessibility document, or LLM prompt. The parser auto-detects the format from the WEBVTT header or the first numeric cue index, then walks every cue while skipping STYLE, REGION, and NOTE blocks that WebVTT permits between the header and the first cue. Each timing line is parsed with both comma-decimal (SRT) and period-decimal (VTT) formats, so a mixed export from a video editor still reads correctly. Inline subtitle markup is recognized and handled: HTML-style styling tags (i, b, u, c, font, ruby, rt, lang), WebVTT voice tags like 'v Speaker', WebVTT timestamp tags like 00:00:01.500 used for karaoke-style word timing, and ASS/SSA override blocks like an8 that sometimes leak into VTT files exported from Aegisub or other editors. Output options give you control over the final shape of the transcript: a paragraph mode (line per cue, merged paragraphs with a configurable pause threshold for the break, or one continuous block of running text), a timestamp mode (none, per cue, or per paragraph), and a speaker-label mode (strip the WebVTT voice tag entirely, keep the speaker name inline as 'Name: ', or break the name out onto its own line above each paragraph). Toggles let you strip every styling tag, deduplicate cues that repeat the same text (common in YouTube auto-captions where the same phrase echoes across two cues), and apply a smart sentence case pass that capitalizes the first letter and adds a trailing period to each paragraph. The output format switch picks plain text or Markdown, where speaker names are bolded and timestamps are italicized for clean rendering inside a docs page, a GitHub README, or a blog post. Live transcript stats show the cue count, paragraph count, word count, character count, total subtitle duration, and an estimated reading time at 220 words per minute, so you know what you are getting before you copy or download. Useful for video producers turning a recording into show notes, course creators converting auto-captions to a clean handout, podcast teams pulling a transcript out of a captioned export, accessibility writers building screen-reader copies of a webinar, translators preparing source text without timing noise, journalists pulling quotes from an interview file, and anyone who needs the words from a subtitle file without the cue indexes and arrows. The whole pipeline runs in your browser, so embargoed recordings, customer interviews, and unreleased course material never leave your device.

Free to use. Works in your browser. No signup, no login.