Licences
A licence is a promise someone else has to keep.
Every row veraeon publishes carries the licence it arrived under, and every licence here is described by what it asks of you — credit, notice, same-terms, no commercial use — rather than by whether it made it onto a list. 50 licences are catalogued. 33 of them are published by default.
Why obligations, and not a list
A list can only say yes or no. Asked about an open-government licence from a country it had never heard of, a list says no — the same no it says to a non-commercial licence that genuinely may not be used. That conflates two very different answers, and the cost of the confusion is paid twice: data that could have been published is thrown away, and the audit trail records that it was thrown away for being unacceptable when it was merely unfamiliar.
So each licence is stored as the set of obligations it places on a user and the set of uses it permits. Publishing is then a policy — "I need to train, and I will not take on non-commercial or no-derivatives terms" — and a licence this registry has not catalogued is recorded as unasserted rather than refused as forbidden. It is still not published. But the corpus can now report how much of itself is unverifiable, which is a number worth having, instead of losing the rows and the number together.
This is not legal advice. The obligations below are read off the licence texts. Whether training a model on a given text creates a derivative work is contested for several of them, so "permits" is deliberately conservative where the answer is argued rather than settled — see no-derivatives below, where the answer is that this registry excludes the use instead of resolving the argument by assertion. Every row carries its licence name so you can run your own policy over it.
The obligations, in full
| Obligation | Means | Detail |
|---|---|---|
| ATTRIBUTION | Credit the source | The licence text and the source must be named wherever the data or a derived work is published. |
| NOTICE_FILE | Carry the notice | The licence text itself has to travel with the data, not just a mention of its name. |
| SHARE_ALIKE | Same terms downstream | Anything derived has to be released under the same licence. This is the obligation most likely to reach a trained model or a derived dataset. |
| DISCLOSE_SOURCE | Say where it came from | The origin of the data has to be stated. |
| STATE_CHANGES | Mark what you changed | Modified copies have to be identified as modified. |
| NETWORK_COPYLEFT | Serving counts as distribution | Offering the work over a network triggers the share-alike obligation (the Affero licences). |
| NO_DERIVATIVES | No modified versions | Modified copies may not be distributed. |
| NON_COMMERCIAL | No commercial use | Commercial use of the data or anything derived from it is not permitted. |
| PATENT_GRANT | Patent licence granted | A benefit rather than a duty: the licensor grants an explicit patent licence. |
| PATENT_RETALIATION | Patent suit ends the licence | If you sue over patents, the licence terminates. |
| TRADEMARK | No trademark rights | The licence grants no rights in names or marks. |
| RESEARCH_ONLY | Research use only | Use is restricted to research and evaluation. |
| UNKNOWN | Terms unknown | No licence is asserted, or the one asserted has not been catalogued. Treated as unverified rather than as permission. |
The catalogue
Every licence the registry knows, grouped by family. The Default column distinguishes what ships under the platform's default policy from what is catalogued and reachable only by naming a policy.
Public domain
| Licence | Family | Obligations | Default |
|---|---|---|---|
| 0BSD BSD Zero Clause License | Public domain | — | default |
| CC-PDDC Creative Commons Public Domain Dedication and Certification | Public domain | — | default |
| CC0-1.0 Creative Commons Zero v1.0 Universal | Public domain | — | default |
| PDDL-1.0 Open Data Commons Public Domain Dedication and Licence | Public domain | — | default |
| Unlicense The Unlicense | Public domain | — | default |
| public-domain Public domain | Public domain | — | default |
Permissive
| Licence | Family | Obligations | Default |
|---|---|---|---|
| Apache-2.0 Apache License 2.0 | Permissive | NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, STATE_CHANGES, TRADEMARK | default |
| BSD-2-Clause BSD 2-Clause "Simplified" License | Permissive | NOTICE_FILE, TRADEMARK | default |
| BSD-3-Clause BSD 3-Clause "New" or "Revised" License | Permissive | NOTICE_FILE, TRADEMARK | default |
| BSL-1.0 Boost Software License 1.0 | Permissive | NOTICE_FILE | default |
| ISC ISC License | Permissive | NOTICE_FILE | default |
| MIT MIT License | Permissive | NOTICE_FILE, TRADEMARK | default |
| MIT-0 MIT No Attribution | Permissive | TRADEMARK | default |
| PostgreSQL PostgreSQL License | Permissive | NOTICE_FILE | default |
| Python-2.0 Python License 2.0 | Permissive | NOTICE_FILE | default |
| X11 X11 License | Permissive | NOTICE_FILE, TRADEMARK | default |
| Zlib zlib License | Permissive | NOTICE_FILE, STATE_CHANGES | default |
Attribution required
| Licence | Family | Obligations | Default |
|---|---|---|---|
| CC-BY-2.5 Creative Commons Attribution 2.5 Generic | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| CC-BY-3.0 Creative Commons Attribution 3.0 Unported | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| CC-BY-4.0 Creative Commons Attribution 4.0 International | Attribution required | ATTRIBUTION, NOTICE_FILE, STATE_CHANGES | default |
| CDLA-Permissive-1.0 Community Data License Agreement – Permissive 1.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| CDLA-Permissive-2.0 Community Data License Agreement – Permissive 2.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| DL-DE-BY-2.0 Datenlizenz Deutschland – Namensnennung 2.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| Etalab-2.0 Licence Ouverte / Open Licence 2.0 (Etalab) | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| KOGL-1.0 Korea Open Government License Type 1 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| NLOD-2.0 Norwegian Licence for Open Government Data 2.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| ODC-By-1.0 Open Data Commons Attribution License 1.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| OGL-UK-2.0 Open Government Licence v2.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| OGL-UK-3.0 Open Government Licence v3.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
| OPL-UK-3.0 Open Parliament Licence v3.0 | Attribution required | ATTRIBUTION, NOTICE_FILE | default |
Share-alike
| Licence | Family | Obligations | Default |
|---|---|---|---|
| CC-BY-SA-3.0 Creative Commons Attribution-ShareAlike 3.0 Unported | Share-alike | ATTRIBUTION, NOTICE_FILE, SHARE_ALIKE | default |
| CC-BY-SA-4.0 Creative Commons Attribution-ShareAlike 4.0 International | Share-alike | ATTRIBUTION, NOTICE_FILE, SHARE_ALIKE, STATE_CHANGES | default |
| ODbL-1.0 Open Data Commons Open Database License 1.0 | Share-alike | ATTRIBUTION, NOTICE_FILE, SHARE_ALIKE | default |
Software copyleft
| Licence | Family | Obligations | Default |
|---|---|---|---|
| AGPL-3.0 GNU Affero General Public License v3.0 | Software copyleft | DISCLOSE_SOURCE, NETWORK_COPYLEFT, NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, SHARE_ALIKE, STATE_CHANGES, TRADEMARK | default |
| EPL-2.0 Eclipse Public License 2.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, SHARE_ALIKE, TRADEMARK | default |
| GPL-1.0 GNU General Public License v1.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, SHARE_ALIKE, STATE_CHANGES | default |
| GPL-2.0 GNU General Public License v2.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, SHARE_ALIKE, STATE_CHANGES, TRADEMARK | default |
| GPL-3.0 GNU General Public License v3.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, SHARE_ALIKE, STATE_CHANGES, TRADEMARK | default |
| LGPL-2.1 GNU Lesser General Public License v2.1 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, SHARE_ALIKE, STATE_CHANGES, TRADEMARK | default |
| LGPL-3.0 GNU Lesser General Public License v3.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, SHARE_ALIKE, STATE_CHANGES, TRADEMARK | default |
| MPL-2.0 Mozilla Public License 2.0 | Software copyleft | DISCLOSE_SOURCE, NOTICE_FILE, PATENT_GRANT, PATENT_RETALIATION, SHARE_ALIKE, TRADEMARK | default |
Non-commercial
| Licence | Family | Obligations | Default |
|---|---|---|---|
| CC-BY-NC-4.0 Creative Commons Attribution-NonCommercial 4.0 International | Non-commercial | ATTRIBUTION, NON_COMMERCIAL, NOTICE_FILE | policy only |
| CC-BY-NC-SA-4.0 Creative Commons Attribution-NonCommercial-ShareAlike 4.0 | Non-commercial | ATTRIBUTION, NON_COMMERCIAL, NOTICE_FILE, SHARE_ALIKE | policy only |
| KOGL-4.0 Korea Open Government License Type 4 | Non-commercial | ATTRIBUTION, NON_COMMERCIAL, NOTICE_FILE | policy only |
No derivatives
| Licence | Family | Obligations | Default |
|---|---|---|---|
| CC-BY-NC-ND-4.0 Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 | No derivatives | ATTRIBUTION, NON_COMMERCIAL, NOTICE_FILE, NO_DERIVATIVES | policy only |
| CC-BY-ND-4.0 Creative Commons Attribution-NoDerivatives 4.0 International | No derivatives | ATTRIBUTION, NOTICE_FILE, NO_DERIVATIVES | policy only |
Restricted / use-limited
| Licence | Family | Obligations | Default |
|---|---|---|---|
| NONE Explicitly no licence | Restricted / use-limited | UNKNOWN | policy only |
| OpenRAIL-M Open RAIL-M License | Restricted / use-limited | ATTRIBUTION, DISCLOSE_SOURCE, NOTICE_FILE, SHARE_ALIKE | default |
| RAIL-1.0 Responsible AI License 1.0 | Restricted / use-limited | ATTRIBUTION, DISCLOSE_SOURCE, NOTICE_FILE, RESEARCH_ONLY | policy only |
Unclassified
| Licence | Family | Obligations | Default |
|---|---|---|---|
| NOASSERTION No licence asserted | Unclassified | UNKNOWN | policy only |
What is published by default
The default publish line is drawn by family: public-domain, permissive, attribution and share-alike. Software copyleft — GPL, AGPL, MPL — is catalogued and reachable, but is not admitted by default, because share-alike reaching a derived program is a stronger claim than share-alike reaching a derived dataset and it is not a claim this platform has made. Non-commercial, no-derivatives, research-only, unasserted and explicitly-unlicensed rows are refused.
Attribution and share-alike rows are not refused. They are published with the obligation recorded per row, and every datasheet counts how many rows carry it — because a release that is 100% licence-covered and 58% attribution-required is two different facts, and only the second one tells you what work you are taking on.
| Policy | What it admits |
|---|---|
| permissive-only | No attribution burden anywhere. For a corpus that cannot carry a credit line. |
| open-dataset (default) | Everything above, plus software copyleft. |
| research-only | Non-commercial and no-derivatives accepted; unasserted still refused. Internal research only. |
| record-only | Everything catalogued, including unasserted rows. For measuring a corpus before deciding what to publish. Not a release policy. |
Contributing under your own licence
If you contribute to veraeon you choose the licence your work goes out under, and the choice is recorded on the row. You do not have to hand over rights to contribute.
| You want | Offer | What it means for us and for you |
|---|---|---|
| Maximum reach, no conditions | CC0-1.0 | Anyone may use it for anything. Nothing to attribute, nothing to track. The default for contributed data when you have no preference. |
| Credit, nothing else | CC-BY-4.0 | Your name travels with the row and every datasheet that includes it counts it. The commonest choice. |
| Credit, and same terms downstream | CC-BY-SA-4.0 | Also published. Derivatives must carry the same licence — including, arguably, models trained on it, which is why the count is reported rather than buried. |
| Code you are contributing | MIT | Contributed tooling and verifiers are accepted under MIT. No notice burden beyond the licence text. |
| Code, with a patent grant | Apache-2.0 | MIT plus an explicit patent licence and a notice file. What veraeon's own tooling ships under. |
| Something else | ask | Open-government licences, national data licences and new Creative Commons ports are added to the registry when someone needs one. Filing an issue is enough; a licence missing from the table is a gap in the table, not a refusal of your work. |
Two things are not negotiable, and they are not licence choices: content that cannot be published at all, and content restricted to research use being placed in a published release. The first is a fixed line. The second is a policy you pick, not one you get to set for everyone else.
Use it from code
The same table is published as JSON, generated from the registry that the gate itself consults, so what you read here is what decides whether a row ships:
curl https://veraeon.org/licences.json | jq '.licences[] | select(.admitted_by_default)
| select(.obligations | index("ATTRIBUTION")) | .spdx'In Python, from the same source of truth:
from veraeon import licences
rec = licences.resolve("CC-BY-SA-4.0")
rec.obligations # frozenset({'ATTRIBUTION', 'NOTICE_FILE', 'SHARE_ALIKE', 'STATE_CHANGES'})
rec.permits # frozenset({'training', 'evaluation', ...})
licences.default_allowed() # the platform's default admit-set
licences.policy_set("permissive-only")