The data behind Vocabulary Collector

Structure of the database

Each dictionary pack is an SQLite 3 file (schema version 1). You can open it with any SQLite tool, for example sqlite3 on the command line.

The database is built for use in the app and changes with it; later versions may add tables or columns. For your own programs rely on pack_info.schema_version and not on a certain number of columns.

Example queries

-- all forms of the word "Haus" (German pack)
SELECT f.surface, f.number, f.grammatical_case
FROM lemma l
JOIN form f ON f.lemma_id = l.id
WHERE l.language_code = 'de' AND l.normalized_lemma = 'haus' AND l.morph_type = 'n'
ORDER BY f.number, f.grammatical_case;

-- translations of a word
SELECT t.language_code, t.translation
FROM lemma l
JOIN lemma_translation t ON t.lemma_id = l.id
WHERE l.language_code = 'de' AND l.normalized_lemma = 'haus';

License and attribution

These data packs contain linguistic data derived from free resources, including UniMorph and Wiktionary. They may include additional normalization, restructuring, filtering, enrichment and error correction.

The data is distributed under the same ShareAlike terms required by the source material: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0). Where parts of the data come from material under CC BY-SA 3.0 (UniMorph), they are distributed under a later version of the same license as that license permits. The license texts: CC BY-SA 4.0, CC BY-SA 3.0.

Sources

• Wiktionary – wiktionary.org. License: CC BY-SA 4.0 for the text of the entries, with additional Wikimedia licensing information as documented by Wiktionary. The data was brought into structured form by the Wiktextract project of Tatu Ylonen (kaikki.org) from the Wiktionary data dumps (English and other editions, 2026).
• UniMorph – unimorph.github.io. License: CC BY-SA 3.0; please also check the license of the specific language dataset in the UniMorph repository.
• Own additions: words, forms and translations entered by hand, forms derived by a rule model (for example the genitive plural in Greek) and translations added with the help of AI services. They are marked as such in the database (form_source, lemma_translation.source_type) and are under the same license.

Changes to the sources

The original data was processed and modified for these datasets, among other things by:

• normalization of forms and features (labels such as case, number, tense),
• restructuring into an own format built for the app,
• removing duplicate entries and cleaning up,
• adding derived information (for example rank, comparison details, reverse direction of translations),
• corrections and consistency checks, partly by hand.

Sharing and attribution

If you share the data or data derived from it, name the sources, link to the license, indicate that you made changes, and share your result under CC BY-SA 4.0. The following text is a suggestion:

Contains data from the Vocabulary Collector data packs (wordcollector.de/en/data-packs/), derived from
Wiktionary (wiktionary.org, CC BY-SA 4.0; processed with Wiktextract/kaikki.org) and
UniMorph (unimorph.github.io, CC BY-SA 3.0). The data was modified.
License: CC BY-SA 4.0 (creativecommons.org/licenses/by-sa/4.0/)

Attribution towards the Wiktionary authors is given by naming Wiktionary with a link to the project and to the revision history of the respective pages (page history at Wiktionary).

Further notes

• The data was extracted by machine from community dictionaries and may contain errors and gaps. There is no warranty for correctness and completeness.
• These datasets and the app are not affiliated with or endorsed by UniMorph, Wiktionary or the Wikimedia Foundation.
• The app itself (program code, interface, texts, logo) is not part of this license; the terms of the App Store apply to it.
• Questions about the license or errors in the data? Write to us through Support & contact.