v2.6.4

New features

• Audiobook mining tab (#71). Queue audiobook and subtitle pairs and mine. Cover art becomes each card's Picture.

• Expression audio on cards (#73). Enable in Settings->Anki and assign a note field.

  • Audio source chain. Expression audio is pulled from an ordered, reorderable list of sources: local audio packs if added, then JapanesePod101, then Google Translate. First hit wins. JapanesePod101+GoogleTTS is the zero-setup default.
  • Local audio pack import. Add a local-audio-yomichan-compatible folder (ajt_japanese, nhk16, forvo_ja, jpod). Indexed in place, so audio files don't move.
  • Google Translate TTS. Free fallback for words no pack or JapanesePod101 covers. Off by default.

Fixed

• i+1 filter checks all unknown lemmas, not just mineable ones (#74).
• Card style presets no longer bleed into unrelated card content.
• Duplicate media filenames no longer re-encoded in batch stores.
• A media file that vanishes mid-run counts as a failure instead of silently succeeding empty.
• A corrupt anki_fields value survives config load instead of crashing.

Repo

• Added the new Anki Miner icon to README.

(Skipped v2.6.3 due to ffmpeg bundling error breaking release.)

Full Changelog: v2.6.2...v2.6.4