Sources & Licenses

Apache-2.0 software; model reports and UD training-data license retained

public domain/no known restrictions with caution for less-than-95-year-old third-party material

CMUdict permissive license

Data.gov catalog marks dataset CC BY and public; field definitions and suppression rules retained

CC BY-SA 4.0 with source assertions retained

Crossref states metadata is openly reusable; abstracts remain excluded unless a compatible item license is verified

DOAJ journal/article metadata is CC0; linked article content remains governed by its item-level license

Use ERIC bibliographic/government-controlled metadata fields; exclude abstracts and full text unless the item carries a separately verified commercial-compatible license or public-domain status

Metadata and explicitly licensed OA subsets only; exclude copyrighted abstracts/full text and all CC BY-NC items from commercial projection

Each dictionary carries its own free/open license in the TEI header; only commercial-compatible dictionaries enter projection with license/source/modification obligations

U.S. public-domain/item notices and Project Gutenberg trademark/terms

discovery metadata; files only through item-level rights allowlist

Official NCES aggregate public data and metadata; preserve release status, definitions, suppression and source attribution

LGPL 2.1 software; bundled rules/data reviewed separately

CC BY 4.0

recordings dedicated to public domain; underlying work and item provenance retained

metadata and item-level rights statements; media filtered per item

Project Gutenberg marks ebook 3201 public domain in the USA

CC BY 4.0

metadata and covers have distinct provenance and rights

discovery only; verify each asset license/provenance

Princeton WordNet License

component licenses in Copyright file must remain traceable

MIT English model listing; exact model package license pinned

User contributions are CC BY-SA; retain author, post URL, revision/license version and share-alike obligations; API throttle and attribution rules apply

Apache-2.0 software; each model training dataset license retained

sentence and audio licenses vary per record; explicit non-PROBLEM license filter and attribution required

Unicode License v3 for data/software files

English EWT CC BY-SA 4.0; other treebanks require independent license filter

MIT software; style packages separate

GPL corpus; copyleft compliance and source offer required

CC BY-SA/GFDL text with revision attribution and share-alike derivatives; non-text media filtered per item

CC0 structured data; linked media rights separate

CC BY-SA 4.0/GFDL text with page-history attribution and modification notice; non-text media require item-level rights filters

Project text reuse follows CC BY-SA/GFDL plus underlying-work public-domain or item-specific status; provenance and page history retained

CC BY-SA/GFDL text with revision attribution and share-alike derivatives; non-text media filtered per item

CC BY-SA and applicable contribution licenses; revision provenance required

Only World Bank-produced Open Data indicators explicitly carrying CC BY 4.0 enter projection; third-party and non-open catalog datasets remain excluded

MIT