Dictionary sources and licences
wordfinder uses one active SQLite index per language. The table shows whether a demonstration or production index is active.
Current status
| Language | Mode | Words | Source | Licence |
|---|---|---|---|---|
| Nederlands | Production index | 592,829 | Wiktionary via Wiktextract en Kaikki.org | CC BY-SA / GFDL |
| English | Production index | 998,365 | Wiktionary via Wiktextract en Kaikki.org | CC BY-SA / GFDL |
| Français | Production index | 1,660,873 | Wiktionary via Wiktextract en Kaikki.org | CC BY-SA / GFDL |
| Deutsch | Production index | 879,008 | Wiktionary via Wiktextract en Kaikki.org | CC BY-SA / GFDL |
Recommended production route
For all four languages, the importer can process machine-readable Wiktionary extracts produced by Wiktextract/Kaikki.org. The data follows Wiktionary licences, including CC BY-SA and GFDL. OpenTaal is an additional Dutch option under the licence choices stated by OpenTaal.
For every import, record the exact dump or release version, source URL, licence, attribution, import date and transformations in data/dictionaries/manifest.json.
Import without changing the design
- Download the chosen dataset directly from its source.
- Run npm run dictionary:import -- --help for all options.
- Build a SQLite index for each language with dictionary:build-index, or migrate version 0.2 with dictionary:migrate-v02.
- Validate the manifest with npm run dictionary:validate.
- Validate and benchmark the index, activate it and restart the Node.js server.
Licence review before publication
A free licence may permit commercial use while still requiring attribution, preservation of licence notices or ShareAlike. Have the final dataset and publication method reviewed. This page is technical documentation, not legal advice.
Kaikki.org raw data · Wikimedia dumps · CC BY-SA 4.0 · OpenTaal