lysandre/anki-words

anki:Kaishi 1.5k Vocabulary Dataset

Japanese vocabulary dataset exported from Anki deck.

Dataset Information

• Deck Name: anki:Kaishi 1.5k
• Word Count: 478
• Format: JSONL (one word per line)
• Last Updated: 2026-02-21

Schema

Each word entry contains:

• word (string): The Japanese word in kanji/kana

• reading (string): The reading in hiragana/katakana

• meaning (string): English translation/meaning

• jlpt: JLPT level (N5, N4, N3, N2, N1)

  • N5: Basic level (easiest)
  • N4: Elementary level
  • N3: Intermediate level
  • N2: Pre-advanced level
  • N1: Advanced level (hardest)

Optional fields (may be added in future updates):

• pos: Part of speech (noun, verb, adjective, etc.)
• tags: List of tags/categories
• frequency_rank: Frequency ranking
• note: Additional notes

Usage

Load with Hugging Face datasets

from datasets import load_dataset

# Load all words
dataset = load_dataset("username/anki-words", split="train")

# Filter by JLPT level (if available)
n5_words = dataset.filter(lambda x: x.get("jlpt") == "N5")

View in dataset viewer

The dataset viewer on Hugging Face Hub will display all words with their readings and meanings.

License

Personal vocabulary dataset for language learning purposes.