license: cc-by-sa-3.0
language:
- id
- jv
- su
- la
- sa
- en
tags:
- linguistics
- etymology
- dictionary
- sqlite
- wiktextract
repo_metadata: true
pretty_name: Kaikki Wiktionary Ultimate SQLite
size_categories:
- 10M<n<100M
Kaikki Ultimate Raw SQLite - World Dictionary Database (2026 Edition)
📌 Overview
This dataset is a high-performance SQLite conversion of the massive Kaikki.org (Wiktextract) raw data. It contains millions of lexical entries across thousands of languages, preserved in its absolute raw JSON format to ensure zero data loss. This database is designed for developers, linguists, and AI researchers who need a structured, indexed, and offline-ready version of the world's most comprehensive dictionary.
🚀 Key Features
- Zero Data Loss: Every single detail from the original raw-wiktextract-data.jsonl is preserved in the full_json column.
- Massive Scale: Approximately 25 GB of structured data covering etymology, senses, pronunciations, and translations.
- Performance Optimized: Includes pre-built B-Tree Indexes on the word and lang columns for millisecond query response times.
- Single File Convenience: No need to parse millions of JSON lines; just connect with sqlite3 and start querying.
📊 Database Schema
The database contains a single table named dict with the following structure:
| Column | Type | Description |
|---|---|---|
| word | TEXT | The headword/entry (Indexed). |
| lang | TEXT | Language name (e.g., "Indonesian", "Latin", "Sanskrit") (Indexed). |
| pos | TEXT | Part of Speech (e.g., "noun", "verb", "adj"). |
| full_json | TEXT | The original raw JSON string containing all metadata. |
🛠 How to Use
You can access this dataset using Python or any SQLite-compatible tool.
Quick Python Example:
import sqlite3
import json
# Connect to the database
conn = sqlite3.connect("kaikki_ultimate_raw.db")
cursor = conn.cursor()
# Search for a word
word_to_find = "surya"
cursor.execute("SELECT full_json FROM dict WHERE word = ?", (word_to_find,))
results = cursor.fetchall()
for row in results:
data = json.loads(row[0])
print(f"Language: {data.get('lang')}")
print(f"Etymology: {data.get('etymology_text')}")
conn.close()
🌐 Target Languages Included
This database is a universal collection, including but not limited to:
- Modern Languages: Indonesian, English, Spanish, Mandarin, etc.
- Regional Languages: Javanese, Sundanese, Minangkabau, Balinese.
- Classical/Root Languages: Latin, Ancient Greek, Sanskrit, Old Javanese, Arabic.
📜 Source & Attribution
The data is sourced from Kaikki.org, which extracts data from Wiktionary. All data is licensed under Creative Commons Attribution-ShareAlike 3.0 (CC BY-SA 3.0).
💡 Potential Use Cases
- Etymology Research: Trace back the roots of Indonesian words to Sanskrit or Latin.
- AI Training: Fine-tune LLMs for translation or morphological analysis.
- App Development: Build powerful offline dictionary apps (e.g., "HafalKuat").
- Linguistic Analysis: Statistical analysis of word formations across different language families.