JapaneseMining

Anki add-on for Japanese sentence mining.

Mine words from real sentences, keep track of which kanji you already know (via Remembering the Kanji / Heisig), fill dictionary data from Jisho, translate example sentences with DeepL, and optionally generate audio with HyperTTS — all from inside Anki.

Features

• Sentence mining workflow

  • Type or paste a Japanese example sentence.
  • Token chips appear above the field (Sudachi). Click a token → it goes into the Word field.
  • One-click DeepL translation into the Translation field.
  • Jisho lookup / quick-fill for reading, meaning, POS, tags, etc.

• Kanji knowledge tracking (RTK / Heisig)

  • Works with a Remembering the Kanji deck.
  • Automatically marks whether every kanji in a mined word is already known.
  • Fills Kanji Keywords and Kanji Meanings.
  • “Add Unknown Kanji” and export of learned kanji.
  • Heatmap + difficult-kanji views.

• Profiles by note type

  • Jisho and Translate settings are keyed by note type.
  • You can use different field mappings for different mining note types.

• Backup

  • Snapshot of the live RTK deck.
  • Restore creates a new deck (safe).

• HyperTTS integration (optional)

  • Configure once; use from the mining workflow.

• Help inside the add-on

  • Quick Start + detailed pages for every major feature.

Requirements

• Anki: Tested on 25.09.4. Should work on recent Anki versions (25.x / 26.x). Older versions are untested.
• Internet connection for Jisho, DeepL, and HyperTTS (if used).
• A DeepL API key if you want automatic translation (the free tier works).

Optional but recommended:

• A Remembering the Kanji (Heisig) deck for kanji tracking
• HyperTTS if you want automatic audio

Quick Start

1. Install the add-on and restart Anki.
2. JapaneseMining → Settings → General
   Create the JapaneseMining note type (keep the default field names).
3. Settings → RTK
   • New to RTK → create the deck + note type.
   • Already have an RTK deck → map it in Deck Mapping, then run Export Learned Kanji.
4. (Optional) Settings → Jisho and Settings → Translate
   Enable and map fields for your note type. Save.
5. Open the Add window, choose the mining note type, and start mining:
   • Write an example sentence → tokens appear → click a token into Word.
   • Translate, look up on Jisho, add the card.
   • Kanji knowledge is updated automatically.

Full walkthrough lives in Settings → Help → Quick Start.

Configuration

All settings live under JapaneseMining → Settings.

Config is stored per Anki profile. The add-on is careful about surviving note-type renames where possible.

Menu actions (JapaneseMining)

• Show Today’s Progress
• Show Kanji (heatmap + difficult kanji)
• Settings
• Soft / Force Update (keywords, meanings, everything)

Privacy & data

• DeepL and Jisho calls go over the network (your API key for DeepL).
• Learned-kanji cache and today’s-progress files live within the add-on.
• No analytics, no telemetry, no “buy me a coffee” prompts.

Known limitations / first release

This is the first public version. Expect rough edges.

• Some edge cases around note-type changes, suspended cards, and multi-profile setups may still surface.
• HyperTTS and certain Jisho fill modes benefit from a bit of manual tuning per note type.

Always back up your Anki collection before installing or updating any add-on.

Bug reports & contact

Found a bug or have feedback? Please get in touch.

When reporting a problem, include:

• Anki version (Help → About)
• JapaneseMining version
• Steps to reproduce
• What you expected vs what actually happened
• Any error message or screenshot if available

Contact options:

• GitHub Issues: https://github.com/programmerare/JapaneseMining/issues
• AnkiWeb support URL (set when you upload the add-on)

I read every report. Even short ones help.

Versioning

This project follows Semantic Versioning (MAJOR.MINOR.PATCH).

• MAJOR – breaking changes
• MINOR – new features (backward-compatible)
• PATCH – bug fixes only

First public release is 1.0.0.

Credits

Built for people who actually mine sentences and use Heisig.

Special thanks to the Anki, Sudachi, Jisho, DeepL, and HyperTTS ecosystems.