Live Mining must also be enabled on the options page, not just configured.
Live Mining requires Anki and the Ankiconnect plugin.
If Live Mining is enabled, then when you mine notes to the mining storage, Nazeka will also attempt to connect to Ankiconnect and tell it to insert a new note.
This is still experimental. Nazeka might not show any error messages if Ankiconnect fails to insert the new note. I only tested errors shown when Anki isn't running.
nothing - Nothing. found_spelling - The spelling you clicked on to mine the word (e.g. "行く") (this is always the "found spelling" described above) readings - Readings (this is the list of readings you see in the mining UI when mining the word) definitions - Definition (HTML; has spans with class names, newlines are <br> tags) definitions_raw - Definition (plaintext, linebreaks are LF characters; only for templates with special CSS or JS) found_text - Text the word was found as (e.g. "行きます"). Also known as "found text". context - The sentence you found the word in. Sentence boundaries are a guess, so it might be cut off funny and need manual editing inside of anki. context_left - The left half of the sentence you found the word in, until the found text. context_right - The right half of the sentence you found the word in, after the found text. audio_anki - anki audio tag (e.g. "[sound:nazeka_audio_ねにもつ_根に持つ.mp3]"), downloaded by Nazeka, sent to anki with Ankiconnect's storeMediaFile command; added even if the storeMediaFile command fails; not all words have audio, may be blank audio_kanji - spelling portion of audio file's name, does no downloading of files. if there are no spellings, it's identical to audio_reading. may be blank. audio_reading - reading portion of audio file's name, does no downloading of files. may be blank. If you use the "audio" field, then the audio file is downloaded from jpod101's servers when you mine the card, and sent to Anki's media folder with Ankiconnect's storeMediaFile action. This only works when you're connected to the internet. The "audio_kanji" and "audio_reading" fields are for people with the JDIC audio pack, or recovering the audio manually if you mined while your internet was down, or for people who want to play the audio file from jpod101's servers directly so they don't need to store it as a media file. (Not sure whether Anki currently supports that, but it should be possible in the future.) jmdict_id - JMDict numeric ID url - URL of the page the card was mined from (if it was the reader page, it will look like jibberish) time_unix - Time when card was mined (unix timestamp, milliseconds, can be converted to a localized date within Anki by using javascript in the card model/template) time_local - Time when card was mined (localized date, depends on browser and OS) readings_raw - The unfiltered list of readings associated with the given word. Tags and restriction information, like さくや not being a reading of 夕べ, are stripped. spellings_raw - The unfiltered list of spellings associated with the given word. Tags and restriction information are stripped.