legal: publish v3 licenses from AWS server

licenses/DATASET_LICENSE.md

@@ -1,76 +1,92 @@
-TokenAI Neo Dataset License v2.0 — Non-Commercial / No-Rebranding
+TokenAI Neo Dataset License v3.0 — Non-Commercial / No-Rebranding
 
 Copyright (c) 2026 TokenAI / Assem Sabry.
 
-TokenAI is a non-profit startup founded in 2025 by Assem Sabry.
-Contact: info@tokenaia.llc. Website: https://tokenai.llc.
+TokenAI is a non-profit startup founded in 2025 by Assem Sabry, based in Alexandria, Egypt. Contact: info@tokenai.llc. Website: https://tokenai.llc.
 
-This license applies to the dataset published at:
-https://huggingface.co/datasets/tokenaii/Neo-dataset.
+This license applies to the dataset published at https://huggingface.co/datasets/tokenaii/Neo-dataset.
 
-## 1. Limited permission
+## 1. Definitions
 
-Permission is granted only for personal, private, non-commercial educational,
-academic, and non-profit research use. Commercial, monetized, paid,
-revenue-generating, sponsored, advertising-supported, client, production,
-business, resale, and financially beneficial use is prohibited.
+**Commercial Use** means any use by or for a for-profit person or organization, or any use connected with a paid product, service, API, subscription, advertising, sponsorship, client work, employment, production business operation, resale, fundraising, investment, or business advantage.
 
-No person or organization may obtain direct or indirect monetary benefit from
-the dataset, a copy, an export, a hosted dataset service, or a product whose
-value depends materially on this dataset.
+**Financial Benefit** includes revenue, fees, salary, commission, donations, sponsorship, investment value, cost savings, lead generation, competitive advantage, in-kind compensation, or any other direct or indirect economic value.
 
-## 2. Mandatory identity, reference, and provenance
+**Derivative Model** means any model or model artifact materially derived from this dataset, including fine-tuned, merged, distilled, quantized, pruned, converted, adapted, LoRA/adapter, checkpoint, embedding, teacher/student, synthetic expansion, or behaviorally substituted model.
 
-Every permitted use must identify the data exactly as:
+**Official Dataset Name** means **TokenAI Neo Decision Model Dataset**. The canonical dataset reference is `tokenaii/Neo-dataset`.
 
-**TokenAI Neo Decision Model Dataset**
+## 2. Limited permission
 
-and must display both canonical references:
+Permission is granted only for personal, private, non-commercial educational, academic, and non-profit research use. No Commercial Use or Financial Benefit is permitted. This license does not grant permission to redistribute the data, operate a dataset service, or publish a Derivative Model.
 
-- Dataset: `tokenaii/Neo-dataset` —
-  https://huggingface.co/datasets/tokenaii/Neo-dataset
+## 3. Mandatory identity and attribution
+
+Every permitted use must identify the data as **TokenAI Neo Decision Model Dataset** and display:
+
+- Dataset: `tokenaii/Neo-dataset` — https://huggingface.co/datasets/tokenaii/Neo-dataset
 - Model/project: `tokenaii/Neo` — https://huggingface.co/tokenaii/Neo
 
-Every model, report, experiment, or derived artifact trained using this data
-must include:
+Every model, report, experiment, or derived artifact trained using this data must include:
 
-> This work was trained using the TokenAI Neo Decision Model Dataset,
-> `tokenaii/Neo-dataset`, owned by TokenAI.
+> This work was trained using the TokenAI Neo Decision Model Dataset, `tokenaii/Neo-dataset`, owned by TokenAI.
 
-The generator name, seed, manifest, split information, attribution, license,
-and provenance notices must remain intact.
+Generator name, seed, manifest, split information, provenance notices, and this license must remain intact where they accompany the data.
 
-## 3. No redistribution or rebranding
+## 4. No redistribution, rebranding, or commercial use
 
 Without prior written permission from TokenAI, you may not:
 
-1. redistribute, rehost, mirror, sublicense, sell, publish, or make available
-   the dataset or any unchanged copy;
-2. publish substantially equivalent exports, reconstructed copies, scraped
-   mirrors, or replacement downloads intended to substitute for the dataset;
-3. rename, rebrand, white-label, fork under another identity, or publish the
-   dataset under another person, company, project, or organization name;
-4. remove or obscure TokenAI, Assem Sabry, the official dataset name, either
-   canonical reference, provenance, or this license;
-5. use the dataset to train a commercial or monetized model, service, product,
-   API, or workflow;
-6. use derived labels, checkpoints, embeddings, or synthetic expansions to
-   evade the no-redistribution or no-commercial-use restrictions.
+1. redistribute, rehost, mirror, sublicense, sell, publish, or make available the dataset or an unchanged or substantially equivalent copy;
+2. publish reconstructed copies, scraped mirrors, or replacement downloads;
+3. rename, rebrand, white-label, or publish the dataset under another identity;
+4. remove or obscure TokenAI, Assem Sabry, the Official Dataset Name, canonical references, provenance, or this license;
+5. use the dataset to train a commercial or monetized model, service, product, API, or workflow;
+6. use labels, checkpoints, embeddings, or synthetic expansions to evade these restrictions or to obtain Commercial Use or Financial Benefit;
+7. use `TokenAI`, `Neo`, or confusingly similar marks as another product or organization's identity.
+
+Publishing a model trained with this dataset never waives the attribution requirement.
+
+## 5. TokenAI and Neo trademark policy
+
+No trademark, trade-name, logo, domain-name, or endorsement rights are granted. Factual nominative reference is allowed only to identify the source and must not imply sponsorship or certification. No confusingly similar mark or domain may be registered or used. Written permission is required for co-branding, white-labeling, or branding beyond attribution.
+
+## 6. Written permission
+
+Request exceptions directly from TokenAI at **info@tokenai.llc** with the material, intended use, distribution plan, duration, and parties involved. Only a written instrument from an authorized TokenAI representative grants an exception.
+
+## 7. Notice-and-takedown
+
+TokenAI may send written notice identifying the material or URL, the alleged breach, the applicable right or license term, and contact or signature. The recipient must stop use, remove or disable access within five business days, preserve evidence, and confirm the correction. TokenAI may request platform or host assistance. A counter-notice may explain the lawful basis for use, but it does not authorize continued Commercial Use. Urgent or repeat breaches may require immediate removal and termination. This procedure does not limit legal remedies.
+
+## 8. Contributions and future CLA
+
+No contribution is accepted unless expressly invited in writing. Contributors must sign a CLA, confirm ownership or authorization, disclose third-party material, and confirm the contribution contains no secrets, unlawful personal data, or incompatible restrictions. The CLA must grant TokenAI a perpetual, worldwide, irrevocable, royalty-free right to use, modify, publish, sublicense, and relicense the contribution. Contributions do not grant redistribution rights to this dataset.
+
+## 9. Third-party material and dependency review
+
+Third-party datasets, software, tokenizer assets, fonts, and other components retain their own licenses. Users must review the dependency and provenance manifest, preserve upstream notices, comply with each license, and verify compatibility before redistribution or deployment. No third-party material is relicensed by this document.
+
+## 10. Patents
+
+No patent license is granted, whether expressly, by implication, estoppel, or otherwise. All patent rights are reserved.
+
+## 11. Termination
+
+Any breach automatically terminates permission. All use and copies must stop and be removed after termination. TokenAI reserves all legal remedies.
+
+## 12. Severability
+
+An invalid provision will be limited or severed only as necessary; the rest of the license remains effective.
 
-Publishing a model trained with this dataset does not waive the mandatory
-dataset attribution and provenance statement.
+## 13. Entire agreement
 
-## 4. Third-party material
+This license, linked notices, and written permissions are the entire agreement for the covered dataset and supersede prior statements about it. Third-party licenses remain separate.
 
-Third-party datasets mentioned in the Neo repository are not included in this
-release and retain their own licenses. This license does not grant rights to
-third-party material.
+## 14. Amendments
 
-## 5. Termination and disclaimer
+TokenAI may publish revised terms for future releases. A release remains under the version shipped with it unless written permission says otherwise. No retroactive expansion or removal of restrictions occurs without written agreement.
 
-Any breach automatically terminates permission. All use must stop and copies
-must be removed after termination.
+## 15. Disclaimer
 
-THE DATASET IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. THIS CUSTOM
-LICENSE IS NOT LEGAL ADVICE AND MUST BE REVIEWED BY QUALIFIED COUNSEL BEFORE
-ENFORCEMENT.
+THE DATASET IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. THIS CUSTOM LICENSE IS NOT LEGAL ADVICE AND SHOULD BE REVIEWED BY QUALIFIED COUNSEL BEFORE ENFORCEMENT.