Turn subtitles into a PDF transcript
Make a readable PDF transcript from an SRT, VTT, ASS or SBV file. Every alphabet renders correctly. Nothing is uploaded to a server.
A subtitle file is already a complete transcript of a film — it is just stored in a shape built for flashing text over video rather than for reading. Translators quoting dialogue, teachers handing out a scene, researchers citing a documentary and anyone studying a language all want the same thing: the words as a document. Subtitle Doctor rebuilds the prose. Line breaks that exist only because a line had to fit the screen are undone, cues that continue a sentence are merged back into paragraphs, and a new paragraph starts where the film goes quiet or a new speaker begins. If you need to cite moments instead, timed mode keeps one entry per cue with its timecode. The PDF is produced by your own browser, which is the reason it handles every writing system correctly.
Worth knowing
- The PDF is made by your browser’s print engine, not by a JavaScript PDF library. That is a deliberate choice: the usual libraries ship Latin-only fonts, so Arabic, Hindi, Thai, Chinese and Cyrillic subtitles come out as empty boxes unless megabytes of font data are embedded per language.
- The document is set in Noto rather than the site’s own typeface, and that is a portability decision, not an aesthetic one: a PDF only travels if its fonts are embedded, and macOS system faces like PingFang and Hiragino set the restricted-embedding flag, so they get referenced by name and turn into blank space on a machine that lacks them. Noto is open-licensed, so the glyphs actually travel inside the file.
- Reading mode undoes subtitle line breaks on purpose. A break inside a cue is a layout decision about screen width, not punctuation, so keeping it would produce a document that reads like a ransom note.
- A paragraph break is inserted wherever the film is silent for longer than the gap you set, which tracks scene changes closely enough to give the document real structure without any understanding of the content.
- Only the scripts you actually used are downloaded or embedded. The font set is served with unicode-range, so an English transcript never fetches a byte of the Chinese, Japanese or Korean faces, and the resulting PDF stays small.
- Cue numbers in timed mode stay matched to the original file even when blank cues are skipped, so a number in the PDF still points at the same entry in the subtitle file.
- Drop several files and they become one document, in order, each starting on a fresh page — a whole series in a single PDF.
FAQ
Why does it open the print dialog instead of just downloading a file?
Because that dialog is where the browser exposes its PDF writer, and that writer is the reason every alphabet renders properly. Choose "Save as PDF" as the destination — on macOS it is in the PDF menu at the bottom left of the dialog, and on Windows and Linux it is in the printer list. It is one extra click in exchange for a document that is correct in every language.
Will my Arabic, Hindi or Chinese subtitles come out correctly?
Yes, and the glyphs travel with the file rather than depending on the reader having the right fonts. Arabic letters join and run right to left, Devanagari conjuncts form, and CJK breaks lines in the right places, because your browser does the text shaping and then embeds the open-licensed Noto faces it used. Tools that build the PDF in JavaScript usually fail here, showing empty boxes, because they ship Latin-only fonts.
What is the difference between reading mode and timed mode?
Reading mode produces prose: cues are merged back into paragraphs and the on-screen line breaks are removed, which is what you want for a handout, a translation or study notes. Timed mode keeps one entry per cue alongside its timecode, which is what you want for quoting a specific moment, checking a translation against the original, or doing a QC pass.
Can I get a plain text file instead?
Yes — the SRT to TXT tool strips the timing and markup and gives you a .txt download directly. Use that when you want the words as data. Use this tool when you want a document a person is going to read or print.