Motion Open Attribution（MO BY）

1. 协议名称与版本号 / Agreement Name and Version

本协议名称为 Motion Open Attribution 协议，版本号为 v1.3。本协议适用于条目详情页明确标注适用本协议的动作捕捉数据。被授权方访问、下载、购买、复制、修改、嵌入、训练或以其他方式使用相关条目，即表示其同意受本协议约束。

This agreement is the Motion Open Attribution Agreement, version v1.3. It applies to motion capture data whose Item Detail Page expressly indicates that this Agreement applies. By accessing, downloading, purchasing, copying, modifying, embedding, training on, or otherwise using the relevant Item, Licensee agrees to be bound by this Agreement.

1.1 权利矩阵 / Rights Matrix

2. 定义 / Definitions

2.1 原始文件 / Original Files

“原始文件”指平台向被授权方提供或允许下载的动作捕捉数据文件及其可直接提取、复制或重新打包的数据形式，包括但不限于 BVH、FBX、GLB、纯骨骼数据、带蒙皮动作数据、动作元数据、压缩包、导出文件、转换格式文件以及平台明确标注为交付内容的其他文件。

“Original Files” means the motion capture data files provided or made available for download by the Platform, and any directly extractable, copyable, or repackaged form of such data, including BVH, FBX, GLB, skeleton-only motion data, skinned motion data, motion metadata, archives, exported files, converted format files, and other files expressly marked by the Platform as deliverables.

2.2 条目 / Item

“条目”指平台上单独展示、标价、授权、下载或引用的一组动作数据内容。每个条目的具体授权协议、预览方式、交付格式与访问级别，以该条目详情页显示的信息为准。

“Item” means a set of motion data separately displayed, priced, licensed, downloaded, or cited on the Platform. The applicable license, preview mode, deliverable format, and access level of each Item are determined by the information displayed on the Item Detail Page.

2.3 被授权方 / Licensee

“被授权方”指按照本协议取得访问、下载、购买、使用或其他授权权益的个人、机构、公司、团队或其他法律主体。被授权方包括其在同一项目中合理需要接触原始文件的雇员、承包商、供应商或项目成员，但该等人员仅可为被授权方项目目的使用原始文件。

“Licensee” means the individual, institution, company, team, or other legal entity that obtains access, download, purchase, use, or other licensed rights under this Agreement. Licensee includes employees, contractors, vendors, or project members who reasonably need access to the Original Files for the same project, provided that such persons may use the Original Files only for Licensee’s project purposes.

2.4 嵌入使用 / Embedded Use

“嵌入使用”指将授权数据作为游戏、影视、动画、广告、软件、交互应用、研究展示、教学内容、机器人演示、可视化作品或其他最终产品的一部分进行整合，使最终用户通常不能将原始文件作为独立素材提取、下载、转售或再授权。

“Embedded Use” means integrating the licensed data into a game, film, animation, advertisement, software, interactive application, research presentation, educational content, robotics demonstration, visualization work, or other End Product, in a manner that ordinarily does not allow end users to extract, download, resell, or sublicense the Original Files as standalone assets.

2.5 再分发 / Redistribution

“再分发”是指以下禁止行为：

• 将原始文件上传至其他内容平台或素材市场；
• 将原始文件打包后单独出售；
• 将原始文件提供给团队/项目成员以外的第三方下载。

以下行为不属于再分发：

• 将数据嵌入游戏、影视、软件等最终产品中；
• 在同一项目团队内部成员之间共享文件；
• 使用数据训练 AI 模型（仅限允许 AI 训练的协议）；
• 在学术论文、研究报告、教学材料、技术博客、会议演示或开源代码仓库中，为说明研究方法、展示研究成果或支持可复现性目的，呈现由授权数据生成的截图、渲染图、可视化、统计图表、动作片段预览或分析结果，且未附带可供第三方直接下载、提取或重建原始文件的完整数据。

本协议所称“可重建原始文件的完整数据”，指足以让第三方无需访问平台即可获得与原始文件实质等同的可用动作数据。展示用途的少量、片段化、低精度或经可视化处理的内容，通常不构成此种完整数据。

“Redistribution” means the following prohibited activities:

• uploading the Original Files to other content platforms or asset marketplaces;
• selling the Original Files separately after packaging or bundling them;
• making the Original Files available for download by third parties outside the team/project members.

The following activities are not Redistribution:

• embedding the data into End Products such as games, films, or software;
• sharing the files among members of the same project team;
• using the data to train AI models, only under agreements that permit AI Training;
• presenting screenshots, rendered images, visualizations, charts, motion clip previews, or analysis results derived from the licensed data in academic papers, research reports, educational materials, technical blogs, conference presentations, or open-source code repositories, for the purpose of illustrating research methods, demonstrating results, or supporting reproducibility, provided that no complete data enabling third parties to directly download, extract, or reconstruct the Original Files is included.

“Complete data from which the Original Files can be reconstructed” refers to data sufficient for a third party to obtain usable motion data substantially equivalent to the Original Files without accessing the Platform. Small, fragmentary, low-resolution, or visualized content used for illustration generally does not constitute such complete data.

2.6 AI 训练 / AI Training

“AI 训练”指将授权数据用于训练、微调、蒸馏、评估、验证、强化学习、模仿学习、行为克隆、运动生成、姿态预测、动作识别、行为预测、机器人控制、具身智能系统、机器学习模型、统计模型或其他可从数据中学习参数、策略、表示、分布或控制规律的系统。除协议明确允许外，AI 训练不被视为普通嵌入使用。

“AI Training” means using the licensed data for training, fine-tuning, distillation, evaluation, validation, reinforcement learning, imitation learning, behavior cloning, motion generation, pose prediction, action recognition, behavior prediction, robotics control, embodied intelligence systems, machine learning models, statistical models, or other systems that learn parameters, policies, representations, distributions, or control patterns from data. Unless expressly permitted by the applicable agreement, AI Training is not treated as ordinary Embedded Use.

2.7 衍生模型 / Derived Models

“衍生模型”指使用授权数据进行 AI 训练后形成的模型、模型权重、策略网络、控制器、生成器、判别器、运动先验、行为策略、嵌入表示、API 服务、推理系统或其他机器学习成果，但不包括原始文件本身或可直接还原原始文件的数据集副本。

“Derived Models” means models, model weights, policy networks, controllers, generators, discriminators, motion priors, behavior policies, embeddings, API services, inference systems, or other machine learning outputs created through AI Training using the licensed data. Derived Models do not include the Original Files themselves or dataset copies from which the Original Files can be directly reconstructed.

2.8 最终产品 / End Products

“最终产品”指被授权方对外发布、交付、销售、展示或内部部署的成品项目，包括但不限于游戏、影视作品、动画、广告、软件、网站、展览、教学材料、研究论文附属材料、机器人演示系统、仿真系统和可视化项目。

“End Products” means finished projects released, delivered, sold, displayed, or internally deployed by Licensee, including games, films, animations, advertisements, software, websites, exhibitions, educational materials, research supplementary materials, robotics demonstration systems, simulation systems, and visualization projects.

2.9 平台展示角色 / Platform Preview Characters

平台用于预览、演示或展示动作数据的角色、模型、贴图、材质、绑定资产或界面元素，如未在条目详情页明确列为可交付内容，不属于本协议授权范围。

Characters, models, textures, materials, rigging assets, or interface elements used by the Platform for preview, demonstration, or display are not licensed under this Agreement unless expressly listed as deliverables on the Item Detail Page.

3. 授权范围 / License Grant

本协议在 Motion Open 的开放授权基础上增加署名义务，适用于希望保留署名权的创作者、学术动作、社区动作和 demo 数据。在被授权方遵守本协议并履行署名义务的前提下，平台与/或创作者授予被授权方一项永久、全球、非独占、不可转让、免版税、免费的使用授权。

被授权方可以为任何商业或非商业目的复制、下载、存储、修改、转换、重定向骨骼、调节速度、裁剪、拼接、清理、适配、导入、渲染、展示、表演、嵌入使用授权数据，并可以将授权数据用于 AI 训练。被授权方可以将授权数据嵌入最终产品，也可以使用授权数据训练衍生模型，并对最终产品或衍生模型进行商业使用、部署、销售、许可、展示或发布。

本协议保留原始文件再分发限制，目的在于保护平台数据库完整性、防止整库搬运和素材平台化倒卖；该限制不影响被授权方对最终产品和衍生模型的正常商业使用。

This Agreement adds an attribution obligation to the open license track and applies to creators who wish to retain attribution, academic motions, community motions, and demo data. Subject to Licensee’s compliance with this Agreement and performance of the attribution obligation, the Platform and/or creator grants Licensee a perpetual, worldwide, non-exclusive, non-transferable, royalty-free, free-of-charge license.

Licensee may copy, download, store, modify, convert, retarget skeletons, adjust speed, trim, splice, clean, adapt, import, render, display, perform, and use the licensed data in Embedded Use for any commercial or non-commercial purpose, and may use the licensed data for AI Training. Licensee may embed the licensed data into End Products, use the licensed data to train Derived Models, and commercially use, deploy, sell, license, display, or release such End Products or Derived Models.

This Agreement retains a restriction on Redistribution of Original Files to protect the integrity of the Platform database and prevent wholesale copying or asset-marketplace resale. This restriction does not affect Licensee’s ordinary commercial use of End Products and Derived Models.

使用规模上限 / Usage Volume Cap

本协议项下的授权设有累计使用规模上限：被授权方对单一授权条目的全场景累计使用不得超过 100,000（十万）次。计量单位包括但不限于：包含该条目的成品分发份数、终端用户数、API 或模型调用次数，或条目说明、平台规则另行约定的等效计量方式。

达到上限后，被授权方应停止新增使用，或购买适用 Motion Commercial 或 Motion Commercial AI 协议的授权以解除规模限制。达到上限前已经发生的合规使用不受影响。

为便于核查，平台保留要求被授权方就其使用规模提供合理说明的权利；本上限属于诚信约束与升级义务，被授权方应在善意基础上评估其使用是否接近或达到上限。

The license under this Agreement is subject to a cumulative usage volume cap: Licensee's total cumulative use of a single licensed Item across all scenarios shall not exceed 100,000 (one hundred thousand) instances. Counting units include, without limitation: units of distributed end products incorporating the Item, end users, API or model invocation counts, or equivalent metrics specified on the Item page or by Platform rules.

Upon reaching the cap, Licensee shall cease additional use or purchase a license under the Motion Commercial or Motion Commercial AI Agreement to remove the volume limit. Compliant use made before reaching the cap is not affected.

For verification purposes, the Platform reserves the right to request reasonable information from Licensee regarding usage volume. This cap operates as a good-faith obligation and an upgrade obligation; Licensee shall assess in good faith whether its use approaches or reaches the cap.

4. 使用条件：署名义务 / Conditions of Use: Attribution

被授权方使用适用本协议的授权数据时，应在合理可行的方式下进行署名。署名应尽量包含以下信息：

• 创作者名称；
• 平台名称：墨影数据（MocaDB）；
• 条目详情页链接或条目名称；
• 协议名称：Motion Open Attribution；
• 如对动作数据进行了显著修改，可注明“已修改”或类似说明。

署名位置可根据使用场景合理调整。例如：

• 游戏、软件、App、网站：可放在 credits、about、文档、帮助页、素材说明、项目页面或许可证页面；
• 影视、动画、广告：可放在片尾、项目说明、制作花絮、官网或宣传资料中；
• 研究、教学、论文、报告：可放在引用、数据来源、附录、补充材料或项目主页中；
• AI 训练：应在训练数据来源声明、模型卡、数据集说明、技术报告、内部数据登记表或类似文档中注明数据来源；不要求衍生模型的每一次推理输出逐条署名；
• 机器人或具身智能产品：可在技术文档、数据来源说明、模型说明、产品合规材料或内部数据台账中署名。

如条目详情页提供了推荐署名格式，被授权方应优先采用该推荐格式；如因版面、保密、产品形态或技术原因无法完整展示，可采用实质等同、合理可见或可追溯的署名方式。

When using licensed data governed by this Agreement, Licensee shall provide attribution in a reasonably practicable manner. Attribution should, where possible, include:

• creator name;
• Platform name: MocaDB;
• Item Detail Page link or Item name;
• license name: Motion Open Attribution;
• if the motion data has been materially modified, a notice such as “modified” or similar wording.

Attribution placement may be reasonably adapted to the use case. For example:

• games, software, apps, and websites: credits, about pages, documentation, help pages, asset notices, project pages, or license pages;
• films, animations, and advertisements: end credits, project descriptions, behind-the-scenes materials, official websites, or promotional materials;
• research, education, papers, and reports: citations, data source sections, appendices, supplementary materials, or project homepages;
• AI Training: data source statements, model cards, dataset documentation, technical reports, internal data registries, or similar documents; attribution is not required for every inference output of a Derived Model;
• robotics or embodied intelligence products: technical documentation, data source statements, model documentation, product compliance materials, or internal data ledgers.

If the Item Detail Page provides a recommended attribution format, Licensee should prefer that format. If complete display is not feasible due to layout, confidentiality, product form, or technical reasons, Licensee may use a substantially equivalent, reasonably visible, or traceable attribution method.

4.1 署名义务的补正与授权终止 / Cure and Termination for Attribution Breach

被授权方未按本条履行署名义务的，应在收到平台或创作者通知后，于合理期限内补正。合理期限可根据使用场景、发布渠道、技术可行性、修改周期和不合规程度确定；被授权方应尽快移除不合规状态或补充合理署名。

经通知后，被授权方仍未在合理期限内补正的，平台或创作者可终止该被授权方就相关授权数据享有的本协议授权。该终止不影响平台或创作者就已经发生的违约、侵权或其他不当使用依法主张权利；但如被授权方已在合理期限内完成补正，则不因该已补正的署名瑕疵当然导致授权终止。

If Licensee fails to comply with the attribution obligations in this Section, Licensee shall cure the failure within a reasonable period after receiving notice from the Platform or the creator. The reasonable period may depend on the use case, publication channel, technical feasibility, update cycle, and degree of non-compliance; Licensee should promptly remove the non-compliant condition or add reasonable attribution.

If Licensee fails to cure the attribution failure within a reasonable period after notice, the Platform or the creator may terminate this Agreement’s license to that Licensee with respect to the relevant licensed data. Such termination does not limit any rights or remedies of the Platform or the creator for prior breach, infringement, or other improper use; however, if Licensee completes the cure within the reasonable period, the cured attribution defect will not by itself automatically terminate the license.

4.2 学术与研究展示 / Academic and Research Presentation

为避免疑义，被授权方为学术研究、教学或成果发表目的，可以：

• 在论文、报告、幻灯片、海报或项目主页中展示由授权数据生成的可视化、截图、图表、动作预览或分析结论；
• 在开源代码仓库中发布使用授权数据的研究代码、分析脚本、实验配置或评测流程；
• 描述、引用、讨论授权数据的特征、统计信息或研究价值。

但被授权方不得在上述场景中附带原始文件本身，或附带可供第三方直接下载、提取或重建原始文件的完整数据、数据集副本或等效打包。研究代码如需引用数据，应通过指向本平台条目详情页的方式提供，而非直接嵌入或附带原始文件。

被授权方对授权数据进行特征提取、统计变换或不可逆处理后形成的研究性衍生数据，如无法用于重建与原始文件实质等同的可用动作数据，可在学术范围内发布；如该衍生数据可被还原为可用的原始动作或等效原始文件，则仍受再分发限制约束。

在论文、项目主页、模型卡、数据说明、开源代码仓库、技术报告或其他学术展示场景中，被授权方仍应按照本协议署名义务提供合理署名。本条不限制本协议明确允许的 AI 训练、衍生模型开发、衍生模型商业部署或模型推理输出使用；但原始训练数据本身及可重建原始文件的完整数据仍不得再分发。

For the avoidance of doubt, for academic research, teaching, or publication purposes, Licensee may:

• present visualizations, screenshots, charts, motion previews, or analytical conclusions derived from the licensed data in papers, reports, slides, posters, or project homepages;
• publish research code, analysis scripts, experiment configurations, or evaluation pipelines that use the licensed data in open-source repositories;
• describe, cite, or discuss the characteristics, statistics, or research value of the licensed data.

However, Licensee may not include, in the above scenarios, the Original Files themselves, or any complete data, dataset copies, or equivalent packaging that would enable third parties to directly download, extract, or reconstruct the Original Files. Where research code needs to reference the data, it should do so by linking to the Platform Item Detail Page, rather than embedding or attaching the Original Files.

Research-oriented derived data created by Licensee through feature extraction, statistical transformation, or irreversible processing of the licensed data may be published within an academic scope, provided it cannot be used to reconstruct usable motion data substantially equivalent to the Original Files. If such derived data can be reverted into usable original motion or equivalent Original Files, it remains subject to the Redistribution restrictions.

In papers, project homepages, model cards, dataset documentation, open-source repositories, technical reports, or other academic presentation scenarios, Licensee must still provide reasonable attribution in accordance with the attribution obligations of this Agreement. This clause does not restrict AI Training, development of Derived Models, commercial deployment of Derived Models, or use of model inference outputs expressly permitted under this Agreement; however, the original training data itself and complete data from which the Original Files can be reconstructed remain prohibited from Redistribution.

5. 限制与禁止行为 / Restrictions and Prohibited Conduct

除非平台、创作者或另行签署的书面协议明确允许，被授权方不得：

• 对原始文件进行再分发；

• 将原始文件作为独立素材、素材包、数据集、训练数据包、动作库、插件资源、模板资源或可下载文件提供给第三方；

• 将原始文件上传至其他内容平台、素材市场、网盘公开链接、代码仓库、数据集平台或可被第三方下载的位置；

• 出售、出租、出借、转让、再授权、质押或以其他方式处分原始文件本身；

• 使用技术手段规避平台的访问控制、下载限制、会员限制、签名链接、预览限制、格式限制、授权记录或支付流程；

• 使用爬虫、批量抓取、镜像、自动化下载或其他方式复制平台数据库中的内容，除非平台另行书面允许；

• 删除、遮盖、伪造或以误导方式替换合理可见的创作者署名、平台来源或条目链接；

• 将授权数据用于违法、侵害他人合法权益、恶意欺骗、仿冒身份、侵犯隐私、规避安全机制或其他明显不当用途；

• 暗示平台或创作者认可、赞助、背书被授权方项目，除非获得明确书面同意。

Unless expressly permitted by the Platform, the creator, or a separate written agreement, Licensee may not:

• redistribute the Original Files;

• provide the Original Files to third parties as standalone assets, asset packs, datasets, training data packages, motion libraries, plugin resources, template resources, or downloadable files;

• upload the Original Files to other content platforms, asset marketplaces, public cloud links, code repositories, dataset platforms, or locations where third parties can download them;

• sell, rent, lend, transfer, sublicense, pledge, or otherwise dispose of the Original Files themselves;

• circumvent the Platform’s access controls, download limitations, membership restrictions, signed links, preview limitations, format limitations, license records, or payment process;

• use crawlers, bulk scraping, mirroring, automated downloads, or other methods to copy content from the Platform database, unless separately authorized in writing by the Platform;

• remove, obscure, falsify, or misleadingly replace reasonably visible creator attribution, Platform source, or Item links;

• use the licensed data for unlawful purposes, infringement of others’ rights, malicious deception, impersonation, privacy invasion, security circumvention, or other clearly improper purposes;

• imply that the Platform or creator endorses, sponsors, or approves Licensee’s project unless expressly agreed in writing.

6. AI 训练条款 / AI Training Terms

6.1 AI 训练授权 / AI Training Permission

在本协议允许 AI 训练的范围内，被授权方可以将授权数据用于训练、微调、评估、验证、强化学习、模仿学习、行为克隆、姿态生成、动作生成、运动控制、动作识别、行为预测、具身智能、机器人控制、仿真系统或其他机器学习用途。除条目详情页或单独书面协议另有约定外，本协议不区分学术、研究、商业或内部研发用途。

Within the scope in which this Agreement permits AI Training, Licensee may use the licensed data for training, fine-tuning, evaluation, validation, reinforcement learning, imitation learning, behavior cloning, pose generation, motion generation, motion control, action recognition, behavior prediction, embodied intelligence, robotics control, simulation systems, or other machine learning purposes. Unless otherwise stated on the Item Detail Page or in a separate written agreement, this Agreement does not distinguish between academic, research, commercial, or internal R&D uses.

6.2 训练数据与衍生模型的权利切断 / Separation of Training Data and Derived Model Rights

• 使用授权数据训练所得的 AI 模型，其推理输出不受本协议限制。

• 训练所得模型的商业部署与销售不受本协议限制。

• 被授权方对训练所得模型拥有完整的商业使用权。

• 原始训练数据本身仍受本协议约束，不可再分发。

• The inference outputs of AI models trained using the licensed data are not restricted by this Agreement.

• Commercial deployment and sale of models trained using the licensed data are not restricted by this Agreement.

• Licensee has full commercial use rights in the models trained using the licensed data.

• The original training data itself remains subject to this Agreement and may not be redistributed.

6.3 原始数据仍受限制 / Original Data Remains Restricted

即使授权数据被用于 AI 训练，被授权方仍不得将原始文件、可直接还原原始文件的数据集、训练数据包、镜像数据集或实质等同的数据副本出售、公开发布、上传到其他平台、提供给第三方下载或作为独立数据产品再授权。

Even when the licensed data is used for AI Training, Licensee may not sell, publicly release, upload to another platform, make available for third-party download, or sublicense as a standalone data product the Original Files, datasets from which the Original Files can be directly reconstructed, training data packages, mirrored datasets, or substantially equivalent data copies.

6.4 AI 训练场景中的署名 / Attribution in AI Training

被授权方使用授权数据进行 AI 训练时，署名义务主要体现在数据来源层面，而非每一次模型推理输出层面。被授权方应在其可合理控制的模型文档、数据集说明、模型卡、研究论文、技术报告、内部数据登记表或对外数据来源说明中保留署名。除非条目详情页或单独书面协议另有约定，本协议不要求模型 API 的每一次响应、机器人每一次动作输出、生成动画的每一帧或每一个终端用户界面均显示署名。

When Licensee uses the licensed data for AI Training, the attribution obligation primarily applies at the data source level, not to every model inference output. Licensee shall retain attribution in model documentation, dataset documentation, model cards, research papers, technical reports, internal data registries, or external data source statements that Licensee reasonably controls. Unless otherwise stated on the Item Detail Page or in a separate written agreement, this Agreement does not require attribution for every model API response, every robot action output, every generated animation frame, or every end-user interface.

7. 创作者权利与下架条款 / Creator Rights and Takedown

• 创作者可申请将内容从平台下架，停止新用户购买/下载。
• 下架不影响已获得授权用户的使用权（已授权用户的权利永久有效）。
• 创作者在上传前须知情同意此条款。

如平台因版权争议、权利瑕疵、合规风险、技术安全、支付异常或其他合理原因暂停、隐藏或下架某一条目，已合法取得授权的用户仍可在本协议范围内继续使用已取得的授权内容，除非法律、法院命令、行政机关要求或双方另有书面约定要求停止使用。平台可在合理范围内提供替代下载、补档、授权记录证明或其他可行的权利确认方式。

• A creator may request that content be removed from the Platform, thereby stopping new purchases/downloads by new users.
• Removal does not affect the rights of users who have already obtained a license; the rights of already licensed users remain permanently valid.
• The creator must knowingly agree to this clause before uploading.

If the Platform suspends, hides, or removes an Item due to copyright disputes, rights defects, compliance risks, technical security, payment anomalies, or other reasonable reasons, users who have lawfully obtained a license may continue to use the licensed content within the scope of this Agreement, unless applicable law, a court order, an administrative requirement, or a separate written agreement requires otherwise. The Platform may provide replacement downloads, file restoration, proof of license records, or other feasible rights-confirmation measures within a reasonable scope.

8. 免责声明与责任限制 / Disclaimer and Limitation of Liability

鉴于本协议项下内容免费提供，平台与创作者不保证授权数据完全符合被授权方的特定项目目的、技术流程、渲染管线、机器人控制系统、研究方法、动画绑定要求、准确性标准、完整性标准或合规需求。授权数据以“现状”和“可用状态”提供，不作任何明示或默示担保，包括但不限于适销性、特定用途适用性、无错误、不中断、无权利瑕疵、动作物理真实性、运动学精度或与任何第三方软件、硬件、模型、算法、数据集、机器人系统兼容的担保。

在法律允许的最大范围内，平台与创作者不对被授权方或第三方因使用、不能使用、修改、转换、嵌入、展示、训练、部署或分发最终产品而产生的间接损失、利润损失、商誉损失、数据损失、业务中断、替代采购成本或特殊、偶然、惩罚性、后果性损害承担责任。

The Platform and the creator do not warrant that the licensed data will meet Licensee’s specific project purposes, technical pipeline, rendering pipeline, robotics control system, research method, animation rigging requirements, accuracy standard, completeness standard, or compliance needs. The licensed data is provided “as is” and “as available” without any express or implied warranties, including warranties of merchantability, fitness for a particular purpose, error-free operation, uninterrupted availability, absence of rights defects, physical realism of motion, kinematic accuracy, or compatibility with any third-party software, hardware, model, algorithm, dataset, or robotics system.

To the maximum extent permitted by law, the Platform and the creator shall not be liable to Licensee or any third party for indirect loss, loss of profits, loss of goodwill, data loss, business interruption, substitute procurement costs, or special, incidental, punitive, or consequential damages arising from use, inability to use, modification, conversion, embedding, display, training, deployment, or distribution of End Products.

9. 协议变更与版本说明 / Changes and Versioning

平台可以不时更新本协议文本，以修正表述、补充解释、适配业务流程、满足合规要求或引入新的授权模式。对于被授权方已经取得授权的条目，除非更新版本对被授权方更有利，或法律、监管要求、法院命令、双方书面协议另有规定，授权取得时适用的协议版本继续适用于该已授权内容。

如条目详情页显示的授权信息与本协议页面存在差异，以条目详情页在授权取得时显示并记录的授权信息为准。平台可保存授权记录、订单记录、下载记录、协议版本号和条目授权快照，用于权利确认与争议处理。

The Platform may update this Agreement from time to time to improve wording, add explanations, adapt business processes, meet compliance requirements, or introduce new licensing models. For Items already licensed by Licensee, the version applicable at the time the license was obtained will continue to apply to the licensed content, unless the updated version is more favorable to Licensee, or applicable law, regulatory requirements, court orders, or a written agreement between the parties provide otherwise.

If the license information displayed on the Item Detail Page differs from this Agreement page, the license information displayed and recorded on the Item Detail Page at the time the license was obtained shall prevail. The Platform may retain license records, order records, download records, license version numbers, and Item license snapshots for rights confirmation and dispute handling.

10. 争议解决与适用法律 / Governing Law and Dispute Resolution

本协议的订立、生效、解释、履行与争议解决，适用平台运营主体所在地法律或平台服务条款中明确约定的适用法律。若平台服务条款、订单条款或双方另行签署的书面合同对适用法律与争议解决方式有明确约定，从其约定。

双方应首先通过友好协商解决因本协议产生或与本协议有关的争议。协商不成的，任何一方可依据平台服务条款、订单条款或单独书面合同中约定的争议解决方式提交有管辖权的法院或仲裁机构处理。

The formation, validity, interpretation, performance, and dispute resolution of this Agreement shall be governed by the laws of the place where the Platform operator is established, or by the governing law expressly specified in the Platform Terms of Service. If the Platform Terms of Service, order terms, or a separate written contract between the parties expressly provides for governing law and dispute resolution, such provisions shall prevail.

The parties shall first attempt to resolve any dispute arising out of or relating to this Agreement through friendly consultation. If no resolution is reached, either party may submit the dispute to the competent court or arbitral institution in accordance with the dispute resolution mechanism set out in the Platform Terms of Service, order terms, or a separate written contract.