v1.9.2

• A file now finds its title's original language when Radarr or Sonarr store
  the library through a symlink the sweep does not follow, for example
  /mnt/storage pointing at /storage. Before, no file matched, so --auto
  never had the arr's corroboration and left every file to an interactive run.

v1.9.0

• Requires Python 3.11 or newer (was 3.8).
• Norwegian (no) now auto-tags under --auto on the same gates as other
  languages. nn, da and sv stay excluded, since whisper confuses them
  with each other and with no.
• Install from GitHub releases: releases/latest/download/audio-lang-tagger.py
  serves the newest stable release. Betas are published as pre-releases, which
  latest skips.
• The README is split into pages under docs/, and SECURITY.md is added.

v1.8.1

First public release.

• A path that does not exist now stops the run. It used to be skipped, so a
  typo or an unmounted share reported a clean library.
• The --auto gates and the coherence floor are covered by the test suite, so
  the thresholds the README documents fail CI if they drift.

One file, no dependencies beyond Python 3.8+, whisper.cpp, ffmpeg and
mkvtoolnix:

curl -O https://raw.githubusercontent.com/polybjorn/audio-lang-tagger/v1.8.1/audio-lang-tagger.py
chmod +x audio-lang-tagger.py
./audio-lang-tagger.py --list /path/to/your/media