Sometimes, Yomitan does not have the audio for a word. Other times, I am reading a manga or a Light Novel and add a new word, but the sentence does not have the sentence audio(duh). Coeroink with コシラズ(Koshikazu) seems like a decent solution, so I decided to automate it.

1. Download Coeroink & install it
2. Download コシラズ & install it in the coeroink speakers folder.

3. Download this script:

   • git clone https://codeberg.org/aeru/anki-gen
4. We are going to need to do some quick settings, you should have Anki and the AnkiConnect plugin

5. In anki-gen.py, edit

   COEROING_ENGINE_PATH = os.path.expanduser("~/downloads/COEIROINK_LINUX_CPU_v.2.7.1/engine/engine")
   COEROINK_URL = "http://localhost:50032/"
   anki_config = {
    "url":"http://localhost:8765/",
    "deck":"日本語::Mine",
    "fields": ["Word Audio:", "Sentence Audio:"],
    "extra": ["is:suspended","-tag:name"],
   }

   • COEROING_ENGINE_PATH Path to your Coeroink install + /engine/engine
   • in anki_config deck to the name of your anki deck
   • in anki_config fields to the fields that have audio (note: for it to properly work you need to have them be named with a space in between eg. your sentence "Sentence" and the audio field for it "Sentence Audio")
   • in anki_config extr to other query options, see Anki Wiki: Searching
6. if you have nix installed, do chmod +x anki_gen.py, otherwise you need python3, and requests and pydub
7. run ./anki-gen generate-all --speaker コシラズ --style ふきげん -o audio