Thursday, 12 May 2022

Convert SRT to WEBVTT subtitles to plain text

subtotxt: Quickly make plain text files from SRT or VTT

View this project on GitHub


Quickly convert a SubRip .srt or WEBVTT .vtt subtitle file to plain text. Removes timestamps and .srt subtitle line numbers.
This was a quick project thrown together for my girlfriend, she’s still learning English and wanted to be able to read subtitles more like a transcript for some trickier language issues (and to understand the jokes in Friends by discussing them with me).

With a spot of feature creep and some encoding detection needs, it evolved into being able to detect character encoding, along with being able to understand both .srt and .vtt formats to save some pre-processing work.

Update July 2022: Version 2.0 is now availalable fixing a bug on Python 3.10.x that prevented it running


Pop the python file somewhere you can reach it and from a command line use:
python C:\Python\ -f
python C:\Python\ -f subtitle.vtt
The script will check which format the subtitle file is (incase of incorrect file extensions), detect the character encoding used then write out a .txt file with the same name as your input. If the output file already exists it will ask for permission to delete and create a new one.

Advanced Usage:

The script has six more arguments you can parse:

  • –utf8 or -8
    Forces the output file to use UTF-8 encoding. This may eliminate character encoding issues if you cannot view the output file. In practice, if you can read the contents of the input subtitle file successfully the output should work without the need to change the encoding.
  • –pause or -p
    Pause the script at the sanity check stage to let you check some stats before continuing, handy if the output is not working.
  • –screen or -s
    Prints the output to the console while writing to the file, may help with debugging failed outputs.
  • –copy or -c
    Copies input to output without change, appends -copy to filename e.g.:, handy to use with –utf8 to quickly change encoding. Might be useful if your video player app cannot understand your original subtitle file encoding.
  • –overwrite or -o
    Skips asking Output file already exists, delete and make a new one? [y/n] and simply deletes the existing output file to create a new one. Ideal for batch processing.
  • –help or -h
    Shows above information.

Required External Modules

  • Send2Trash Python module to safely delete the old output file on both Win and *nix based systems.
  • cchardet Python module to detect your subtitle file encoding. (Removed for v2.0 release due to issues with Python 3.10.x installs, still used in v1.0 and will work on Python 3.9.x installs)
  • charset_normalizer Python module to detect your subtitle file encoding (v2.0+ supports Python 3.9.x and 3.10.x).

If your system does not these installed, it will auto install them on first use.


  • Fast (aside from initial missing modules install on slow net connections)
  • Input files character encoding formats are autodetected if supported by cchardet [v1.0] or charset_normalizer [v2.0+]
  • Output files are wrote in the same encoding as the input or can be forced to UTF8
  • Should be cross platform friendly thanks to PathLib and Send2Trash
  • Handles UNC style \\myserver\myshare\ paths thanks to PathLib
  • Handles SRT to TXT or WEBVTT to TXT
  • Handles multi line subtitles and subtitle lines with just numbers (does not confuse them with SRT line numbers)
  • WEBVTT: Removes ‘WEBVTT’, ‘Kind: xxxx’, ‘Language: xxx’ headers and Timestamps from output
  • SRT: Removes subtitle line #'s and Timestamps, will not work if first subtitle is not 1 or if duplicated line numbers are present (rare cases but possible), use SubtitleEdit to renumber lines for now if this happens.



Kind: captions
Language: en

00:00:18.590 --> 00:00:21.389
you'll hear a telephone conversation

00:00:21.389 --> 00:00:23.310
now you have some time to look at

00:00:23.310 --> 00:00:27.589
questions one to six

or SRT Input:

00:00:18,590 --> 00:00:21,389
you'll hear a telephone conversation

00:00:21,389 --> 00:00:23,310
now you have some time to look at

00:00:23,310 --> 00:00:27,589
questions one to six


you'll hear a telephone conversation
now you have some time to look at
questions one to six

Examples with non latin characters:

These are random examples take from an SRT website. cchardet detects the encoding as UTF-8-SIG, Notepad++ detects as UTF-8-BOM, these are technically the same thing.

Arabic SRT in UTF-8-BOM / UTF-8-SIG encoding:

00:00:02,425 --> 00:00:20,776
تـرجـمـة وتـعـديـل
الدكتور علي طلال 

00:00:58,425 --> 00:00:59,776

00:01:01,705 --> 00:01:03,462
هل تريدين أنّ تأكلين مجددًا؟

Output with forced UTF-8 encoding:

تـرجـمـة وتـعـديـل
الدكتور علي طلال 
هل تريدين أنّ تأكلين مجددًا؟

Chinese (Simplified) SRT in UTF-8-BOM / UTF-8-SIG encoding:

00:00:58,016 --> 00:00:59,476

00:01:01,270 --> 00:01:03,272

00:01:04,313 --> 00:01:07,276

00:01:09,528 --> 00:01:10,612

Output file in original UTF-8-BOM / UTF-8-SIG encoding:


Future plans:

  • Possibly handle more formats (.ssa Sub Station Alpha would be the other major one I could think of), for now you can use something like SubtitleEdit to convert most other formats to .srt or .vtt. If you have a format you would like to convert to txt, contact me or raise an issue to see if I can add support.
  • GUI option for simple drag and drop usage.
  • Figure out a checking method for misnumbered or duplicate numbered SRT line numbers.
  • Handle stripping out SRT formatting tags for bold, italic etc…


Released as CC0, use it how you wish. If you do use it elsewhere, please be awesome and tag me as the original author. 🙂

No comments:

Post a Comment