The following python tool can be used to let it analyse the audio and subtitle and synchronise the subs to generate a new synced file. It works for offsets at the beginning, not for sync issues later in the file.

https://github.com/smacke/ffsubsync

Install on Linux Mint 22+:

pipx install git+https://github.com/smacke/ffsubsync@latest

Usage:

ffs video_file.mkv -i orig_sub.srt -o new_synced_sub.srt

Untested alternative without limitations:
https://github.com/kaegi/alass

Other alternative:
https://snapcraft.io/subsync
https://sc0ty.pl/
https://subsync.online/
https://github.com/sc0ty/subsync?tab=readme-ov-file

https://github.com/denizsafak/AutoSubSync

JSON-RPC
Method: Input.ExecuteAction
Parameter: {“action”:“subtitledelayminus”}
or
Parameter: {“action”:“subtitledelayplus”}

Extract subtitle with specific metadata title:

for f in *.mkv; do ffmpeg -hide_banner -i "$f" -c copy -map 0:s:m:title:"English [ForcedNarrative]" "${f%.*}.en.forced.srt"; done
for f in *.mkv; do ffmpeg -hide_banner -i "$f" -c copy -map 0:s:m:title:"angličtina" "${f%.*}.en.srt"; done

#extract 2 at once
for f in *.mkv; do ffmpeg -hide_banner -i "$f" -c copy -map 0:s:m:title:"English [ForcedNarrative]" "${f%.*}.en.forced.srt" -c copy -map 0:s:m:title:"angličtina" "${f%.*}.en.srt"; done

This copies the video and audio stream of an mp4 file, sets the first mov_text subtitle stream to 'forced' and the second to normal:

ffmpeg -i input.mp4 -vcodec copy -acodec copy -map 0:v:0 -map 0:a:0 -map 0:s:0 -c:s mov_text -disposition:s:0 forced -map 0:s:1 -disposition:s:1 0 output.mp4

For mkv files, this probably needs to be srt instead of mov_text.