Foundation beta · source record
Sources & licences
The 15,000 foundation-beta entries are machine-assisted candidates covering 1,500 linked concepts in ten languages. They have passed structural, provenance and collision checks, but have not yet received human linguistic review. Labels, broad IPA and definitions may therefore need correction. Audio is not included.
Concepts and definitions
English sense anchors and the original English source definitions come from Open English WordNet 2024, licensed under CC BY 4.0. Language-specific explanations are introduced as project-owned editorial drafts and remain explicitly pending human review.
Sense-to-item alignment, multilingual labels and gap-filling short descriptions come from Wikidata under CC0. A Wikidata description is fetched only when a linked entry lacks a project or reviewed definition; it never replaces one. Wikipedia articles are fetched only when you open the reader, are not stored in the catalogue, and are attributed in that reader under CC BY-SA 4.0.
Entry media
Media suggestions are fetched from Wikimedia Commons only when an entry is opened. The gallery accepts results whose metadata identifies a public-domain, CC0, CC BY or CC BY-SA licence. Every image links to its Commons description page, where the creator, complete licence terms and source record remain available. Results with non-commercial, no-derivatives or unclear licensing are excluded. Opening an entry therefore makes requests to Wikimedia infrastructure.
English pronunciation
Broad pronunciation candidates use CMU Pronouncing Dictionary. The project snapshot is covered by its permissive licence.
SHA-256: 81917843c7f44ce2b094ac63873c2c7a4cf802040792c455ba3ca406891c3d22
Seven IPA streams
German, Turkish, Italian, Spanish, French, Russian and Standard Arabic broad IPA candidates were generated with Epitran 1.25 under the MIT licence. Unvocalized Arabic and Russian stress are especially important human-review targets.
Japanese readings
Japanese reading candidates are derived from the JMdict/EDICT Dictionary Project, used under CC BY-SA 4.0. See the EDRDG licence statement.
SHA-256: 4cf42bac4f5b5fbbc79a56ed71a21646448021d73dcdfe89450021a7c2d8c42d
Standard Chinese readings
Standard Chinese Pinyin candidates are derived from CC-CEDICT, used under CC BY-SA 4.0.
SHA-256: fefb4cbccd1d3dbcd9064f62a27cd045e86579c8bc900a20b4ab69b7f06e7dd8
What “foundation beta” means
Every entry remains marked needs review. Stable entry identifiers let us add corrected labels, reviewed IPA and licensed audio step by step without breaking links. A structural pass is not a claim of linguistic or medical-grade accuracy.