# CausalGPT-RL AWS Marketplace Product-Specific Terms

These CausalGPT-RL AWS Marketplace Product-Specific Terms (the “Product-Specific Terms”) apply to the CausalGPT-RL product made available by CCNets, Inc. (“CCNets”) through AWS Marketplace.

## 1. Relationship with SCMP / EULA

These Product-Specific Terms supplement the applicable AWS Marketplace agreement, the AWS Standard Contract for AWS Marketplace (“SCMP”), or any other applicable end-user license agreement (“EULA”) accepted by Customer in connection with Customer’s subscription to or use of the CausalGPT-RL product through AWS Marketplace.

Except as expressly provided in these Product-Specific Terms, the SCMP or applicable EULA remains unchanged and continues to apply to Customer’s use of the CausalGPT-RL AWS Marketplace product.

In the event of a conflict between these Product-Specific Terms and the SCMP or applicable EULA, these Product-Specific Terms will control solely with respect to CausalGPT-RL product-specific matters.

## 2. Definitions

“Product” means the CausalGPT-RL product made available by CCNets through AWS Marketplace.

“Public Materials” means software, public inference runtime code, model loading utilities, examples, notebooks, documentation, public model bundles, public weights, public checkpoints, or other related materials made available by CCNets through GitHub, PyPI, Hugging Face, public S3 locations, public container registries, or other public repositories or public distribution channels.

“Customer-Specific Model Outputs” means model weights, checkpoints, bundles, snapshots, configuration files, or similar model artifacts generated or provided specifically for Customer through the Product using data provided by Customer or data that Customer is authorized to use.

## 3. Separation of Public Materials and AWS Marketplace Outputs

Public Materials are separate from Customer-Specific Model Outputs generated or provided through the Product.

Public Materials are governed by the license terms applicable to those Public Materials. Customer’s subscription to or purchase of the Product through AWS Marketplace does not modify, expand, replace, or supersede the license terms applicable to Public Materials.

Unless separately stated in writing by CCNets, Public Materials remain subject to any non-commercial, research, evaluation, or other restrictions set forth in the applicable public license.

For clarity, any commercial use rights granted to Customer under the SCMP, applicable EULA, or these Product-Specific Terms apply only to the Product and Customer-Specific Model Outputs, and do not grant commercial use rights to Public Materials.

For further clarity, the public availability of inference runtime code, model loading utilities, examples, or notebooks as Public Materials does not grant or extend to such Public Materials any commercial use rights granted for the AWS Marketplace Product or Customer-Specific Model Outputs.

## 4. Permitted Use of Customer-Specific Model Outputs

Subject to the SCMP or applicable EULA and these Product-Specific Terms, Customer may use Customer-Specific Model Outputs generated or provided through the Product for Customer’s internal production inference and internal use in Customer’s own products or services.

Customer may commercially use results, predictions, action outputs, evaluation outputs, or similar outputs generated through permitted inference use of Customer-Specific Model Outputs in connection with Customer’s internal business operations, products, or services.

This permission applies only to Customer-Specific Model Outputs generated or provided specifically for Customer through the Product.

For clarity, Customer Data and Customer Materials remain owned by Customer as set forth in the SCMP or applicable EULA. Customer-Specific Model Outputs are licensed for use only within the scope expressly permitted by the SCMP or applicable EULA and these Product-Specific Terms.

## 5. No Transfer of Training Algorithm or Core Intellectual Property

For clarity, the CausalGPT-RL training algorithm, training pipeline, source code, model generation method, internal weights generation method, technical know-how, trade secrets, patents, patent applications, inventions, methods, systems, architectures, and related intellectual property rights remain the property of CCNets or its licensors.

Except for the express rights to use the Product and Customer-Specific Model Outputs as permitted under the SCMP or applicable EULA and these Product-Specific Terms, Customer does not acquire any ownership rights, assignment rights, sublicense rights, or separate license rights in or to the foregoing. Such rights may be protected by copyright, patent, trade secret, or other intellectual property laws.

Except to the extent such restriction is prohibited by applicable law, Customer may not extract, reconstruct, copy, reverse engineer, decompile, disassemble, or otherwise attempt to obtain the source code, training algorithm, training method, model generation method, internal structure, or underlying technology of CausalGPT-RL.

## 6. No Resale, Redistribution, Public Upload, or Standalone Runtime Use

For clarity, Customer-Specific Model Outputs, model weights, checkpoints, bundles, snapshots, and any runtime, container, or related components provided through the Product may not be sold, sublicensed, redistributed, publicly uploaded, disclosed, transferred, or otherwise made available to any third party except as expressly permitted under the SCMP or applicable EULA.

Customer may not make Customer-Specific Model Outputs or the CausalGPT-RL runtime, container, or related components available in any of the following forms:

- standalone model;
- public model repository upload;
- public GitHub, Hugging Face, PyPI, S3, container registry, or similar public distribution;
- SDK or runtime product;
- inference platform;
- SaaS product;
- service bureau, application service provider, or managed service offering for third parties; or
- competing machine learning, reinforcement learning, model training, or inference service.

Except as expressly permitted under the SCMP or applicable EULA, Customer may not use the Product, Customer-Specific Model Outputs, runtime components, or container components to provide model training, model generation, inference hosting, reinforcement learning services, or similar services to third parties.

## 7. Checkpoints and Bundles

Checkpoints generated through the Product may be provided for training continuation, retraining, recovery, or related internal training workflows.

Bundles, snapshots, and related artifacts generated through the Product may be used for permitted inference, evaluation, and deployment within Customer’s own internal systems, products, or services, subject to the SCMP or applicable EULA and these Product-Specific Terms.

Nothing in this section grants Customer any right to sell, sublicense, redistribute, publicly upload, disclose, transfer, or otherwise make available to any third party any checkpoint, bundle, snapshot, model weight, runtime component, or container component except as expressly permitted under the SCMP or applicable EULA.

## 8. Reservation of Rights

CCNets reserves all rights not expressly granted to Customer under these Product-Specific Terms.

No rights are granted by implication, estoppel, or otherwise. These Product-Specific Terms do not grant Customer any right to use the name, trademarks, service marks, logos, or branding of CCNets except as expressly permitted in writing by CCNets.

## 9. Confirmation of Relationship with SCMP / EULA

These Product-Specific Terms are intended to clarify the following CausalGPT-RL product-specific matters:

- separation of Public Materials, public inference runtime, and AWS Marketplace outputs;
- permitted use of Customer-Specific Model Outputs;
- no transfer of the training algorithm or related intellectual property rights; and
- prohibition on resale, redistribution, public upload, and standalone runtime or service use.

All other terms, restrictions, disclaimers, remedies, and obligations under the SCMP or applicable EULA remain unchanged.