Suptext is a PGS/SUP to SRT subtitles converter written in Go and uses Tesseract for OCR.

Suptext expects an english PGS/SUP encoded subtitles file as input, and outputs an SRT SubRip file.
If you're unable or couldn't be bothered to extract the SUP file yourself from a video then jump to the Run via Docker section for a utility that supports video input and does that for you.

• Install Tesseract-OCR based on your OS and download the english language package
• Run go install github.com/eliaonceagain/suptext@latest

The following instructions use eliaonceagain/suptext Docker image. The image contains a utility command that extracts eng-PGS/SUP subtitles from a video file using ffmpeg then calls Suptext Go package to decode and output the result SRT SubRip file.

• Run a container based on eliaonceagain/suptext Docker image and mount a media directory.
  If your media is not in the current directory then change "$(pwd)" accordingly.

docker run --rm -itd \
       --name suptext \
       -v "$(pwd)":/mymedia \
       eliaonceagain/suptext:latest

• Verify container is up

docker exec suptext supcli --help

Usage: /usr/bin/supcli [OPTION] <path>
Exactly one of the following options must be set
  -v,  --video <path>      : Set the video file
  -s,  --sup <path>        : Set the supplementary file
  -h,  --help              : Display this help message

• Run supcli command in the container while providing either a video or a PGS/SUP file. Output filename will be identical to input filename besides having .srt extension; e.g. video.mkv -> video.srt

# from video file
docker exec suptext supcli -v "/mymedia/video.mkv"

# or from sup file
docker exec suptext supcli -s "/mymedia/subtitles.sup"

• Example output

docker exec suptext supcli -v "/mymedia/SV-S04E10.mkv"

2024/07/04 17:20:15 Processing video file: SV-S04E10.mkv
2024/07/04 17:20:16 Found eng-PGS subtitles at stream index: 3
2024/07/04 17:20:23 Successfully extracted subtitles from input video
2024/07/04 17:20:23 Reading SUP file: /tmp/tmp.xY7TPdfhCp.sup
2024/07/04 17:20:23 Writing SRT file: /tmp/tmp.xY7TPdfhCp.srt
2024/07/04 17:26:31 Success

v0.1.1 - Added support for fragmented ODS
v0.1.2 - Ignore truncated PCS extension
v0.1.3 - Fix for issue #5 (output contains timestamps but no text)

Not working as expected? Open an issue, add description, and upload a link to the video or subtitles.

Want to further cleanup your output SRT file? Check out SRT-Link, a CLI tool for filtering SRT subtitles based on specific text, characters, or brackets.