成果的权利与用数据训练

工具已经做出东西之后，人会问自己四个问题：做出来的东西归谁所有、能不能拿去交易；模型会不会拿我上传的内容做训练；能不能拒绝这一点；服务会不会给 AI 的作品加标注。

这里是 51 个工具的答复——正如各服务自己所写，并附文档链接。

这不是法律意见，也不是解释。 如果文件本身没有把许可归结为«可以/不可以»，我们也不会这样归结，更不会替你决定某个表述意味着什么。

成果归谁所有

这里关键的不是服务，而是套餐。 在 44 条答复中，有 27 条明确指出权利取决于套餐。人们先免费试用，做完作品——事后才得知它不能出售。

权利取决于套餐 — 27

Auphonic. 权利留在用户手中：«you fully retain any rights that you have with regards to your Content, including copyright»（原文）。条款中没有对成果商业使用的限制，权利上也未声明资费差别。免费资费的实际差别不是法律上的而是技术上的 —— 输出中会嵌入 Auphonic 的片头音。禁止用户把该服务、其输出及衍生物用于 «to develop, train, evaluate, benchmark, or improve any model, algorithm, or system»（原文），也禁止在此基础上构建竞争性产品来源, 已核查 2026-08-07

Bria. 按套餐不同，并且写在条款中，原文为：«Certain Services, including but not limited to API access, advanced AI model features, and commercial use rights, are available exclusively to users with active paid subscriptions or valid licensing agreements»。价格页为 Enterprise 补充了原文的«Commercial use rights for open-source models»和«Unlimited IP… 来源, 已核查 2026-08-07

Cleanvoice AI. 成果权利留在用户手中。Terms & Conditions 第 7.3 条原文：«you also retain all right, title and interest in and to the Results generated through the Services»。给该服务的许可（第 7.4 条）仅为提供所订购的服务而授予。已读内容中没有对成果商业使用的限制。条款和资费中都没有资费差别：定价页上直接写着 «All plans include all features» 来源, 已核查 2026-08-07

Elai.io. 各来源不一致：这里存在产品与条款之间的矛盾，必须直说。服务条款第 2a 条：«Elai grants you a non-transferable, non-exclusive, revocable, limited license to access and use the Website and Services for you own personal, non-commercial use»，以及后文 «You shall not license, sell, rent, lease, transfer, assign, distribute, host, or otherwise commercially exploit the Website or Services… 来源, 已核查 2026-08-07

Fliki. 因资费而异，这是关键。条款规定：用户«retain any and all of Your rights to any Content You create, post, or display on or through the Service»（原文）。但商业许可是付费功能：定价页面的 FAQ 中写道«with our subscription plans you get a commercial license to all the content you create. You can use the content created to publish on Youtube or other social media and monetize it… 来源, 已核查 2026-08-07

getimg.ai. 按资费不同：原文 «Can I use created images for commercial projects? Yes, but only with our paid plans. Images generated under these plans can be used commercially, adhering to the CreativeML Open RAIL-M license»；关于上传内容，条款原文写着 «All materials uploaded remain yours» 来源, 已核查 2026-08-07

HeyGen.【因套餐而异】，而且差异悬殊。付费档：«you own all rights in your User Input or User Output»、«HeyGen does not claim to own any of Your Content»，明确允许商用。免费档：只授予 «limited, non-exclusive, non-transferable, revocable license» 供个人非商业使用，且原文写明 «User Output generated under a Free Plan may not be sold, sublicensed, redistributed, monetized, or used in connection… 来源, 已核查 2026-08-07

HyperWrite. 条款没有直接分配生成文本的权利；使用平台的许可被授予 «for your personal, non-commercial purposes» —— 原文如此，也就是说这份文件并未允许商用。按资费而言，条款文本中没有差别来源, 已核查 2026-08-07

Koala AI. 结果转交用户，原文为：«Furthermore, we hereby assign to you all our right, title, and interest in and to the Content»，其中 Content 指输入和输出的合称；同处还有 «You retain all rights to your Input»。该工具保留将 Content 用于服务运行、遵守法律和执行自身政策的权利。按资费无差异 —— 条款中没有针对个别方案的保留条款来源, 已核查 2026-08-07

Krea. 条款没有直接称用户为所创作内容的所有者。写的是：«Krea may allow you to commercialize certain Content generated from the Services, subject to your adherence to our copyright dispute policy and any other restrictions or requirements (including pricing changes) imposed by Krea»。同时，创作结果落入«User Submission»的定义（«Anything you post, upload, share, store, or otherwise… 来源, 已核查 2026-08-04

LALAL.AI. 条款没有以直接措辞把成果权利转让给用户，也没有描述资费之间的差别。只写了：«You are solely responsible for the usage and distribution of uploaded audio files and end-result audio files»（原文）。同一文件中有一条一般性禁令：«You agree not to reproduce, duplicate, copy, sell, trade, resell or exploit for any commercial purposes, any portion of the Service, use of the Service, or access to the… 来源, 已核查 2026-08-07

Let's Enhance. 条款未直接规定对结果的权利；关于上传内容，原文写道«WE CLAIM NO OWNERSHIP RIGHTS OVER YOUR CONTENT»。条款中各套餐在商业使用上没有差别，但实际存在差别：免费档的结果带水印（价格页）来源, 已核查 2026-08-07

Mubert. 权利留在该服务手中，而许可的范围因资费而大不相同。帮助原文：«All the copyright is owned by Mubert Inc. We license tracks for different use cases»。许可页面：«Mubert owns all the rights to the tracks generated. Our Sync License allows you to use music in videos, podcasts, apps and games»。分档情况：免费的 Ambassador —— 仅限个人项目，不得有广告投放和变现，且必须署名（附 Mubert Render 链接和 #mubert… 来源, 已核查 2026-08-07

Murf AI. 按资费不同。任何付费资费都给予商业权利：«Yes, all paid plans offer commercial rights over the voiceovers generated in the Studio»（原文），包括在 YouTube 和其他流媒体平台上变现。免费试用禁止下载，因此谈不上商业使用。权利是永久的：«commercial rights and business license rights remain valid even after your subscription ends»（原文）。服务条款：«You can use Murf created voices for commercial… 来源, 已核查 2026-08-07

Pika. 定价页在全部四档资费（包括免费的 Basic）上都标注 «Commercial use»，也就是说按资费表商业使用并未加以区分。生成的视频究竟归谁所有，定价页没有说明，而服务条款的正文未能读取来源, 已核查 2026-08-07

Pixelcut. 按套餐不同：原文的«Commercial license»一行出现在 Pro 和 Business 套餐中，免费档没有；使用条款完全没有讨论对结果的权利——那里只有授予工具处理上传内容的非独占许可来源, 已核查 2026-08-07

Playground. 各来源不一致：来源之间存在分歧。价格页区分套餐：免费档为«No royalty-free license»，付费档为原文的«World-wide, royalty free license»。而条款不按套餐区分，对所有人写道：原文为«you own all Assets you create with the Services and (iii) we hereby assign to you all rights, title and interest in and to such Assets for your personal or commercial use» 来源, 已核查 2026-08-07

Postwise. 内容的权利仍归用户，原文为：«You retain any and all of your rights to any Content you submit, post or display on or through the Service»；发布行为赋予该工具在用户账户范围内使用、展示和分发该内容的权利。条款中没有单独关于 AI 输出权利的条款。按资费无差异来源, 已核查 2026-08-07

ProWritingAid. 权利仍归用户，条款原文为：«You retain ownership of all content you post, upload to, or otherwise share on the site»，隐私政策原文为：«We do not retain any rights to any of the documents or other content uploaded into our system»。按资费无差异 —— 文件中没有针对 Free 或 Premium 的保留条款来源, 已核查 2026-08-07

Recraft. 条款区分套餐。免费档：«Assets generated by Free Tier users are Free Tier Assets»，许可为«for your personal use»，并附带限制«(i) no commercial use of Free Tier Assets is permitted; (ii) Free Tier Assets are owned by Recraft and may not be sold, licensed or transferred; and (iii) Free Tier Assets may not be used to train artificial intelligence… 来源, 已核查 2026-08-04

Resemble AI. 条款并未直接把生成音频的权利转让给用户：文中没有关于 Output 归属的表述。Resemble 保留对 Aggregated Data —— 去标识化的衍生数据和元数据 —— 的权利。禁止用户把输出用于 «to train, improve, or otherwise further develop your own or any third party's product»（原文）。条款中未描述资费之间的差别。另一条线是开源模型：Chatterbox Multilingual 声明 «Full model weights on Hugging Face ... No vendor lock-in, no commercial restrictions»，采用 MIT… 来源, 已核查 2026-08-07

SOUNDRAW. 条款直接把权利留给工具方：«All Intellectual Property Rights pertaining to the Service and the programs that comprise the Service belong to the Company. The neighboring rights of producers of phonograms to the sound sources used in the Generated Music belong to the Company. The Company shall grant the Registered User a license to use the Generated… 来源, 已核查 2026-08-04

Suno. 条款按资费划分权利。免费：«you covenant and agree that you will only use Outputs generated from Submissions made by you through the Service solely for your lawful, internal, personal and non-commercial purposes»。付费：«Suno hereby assigns to you all of its right, title and interest in and to any Output owned by Suno and generated from Submissions made… 来源, 已核查 2026-08-04

Taplio. 内容的权利仍归用户，原文为：«You retain any and all of your rights to any Content you submit, post or display on or through Service»；同时，通过本服务发布内容即赋予公司 «use, modify, publicly perform, publicly display, reproduce, and distribute such Content on and through Service» 的权利，其中包括向该服务的其他用户展示。条款中没有关于 AI 生成文本权利的直接条款。按资费无差异来源, 已核查 2026-08-07

Topaz Labs. 按套餐不同，而且直接写在价格表中：Topaz Studio 为«limited commercial use»，Topaz Studio Pro 和 Topaz Image Web Pro 为原文的«Full commercial use»。处理后的图像归谁所有，docs.topazlabs.com 上的条款未作规定来源, 已核查 2026-08-07

Upscale.media. 条款未规定对结果的权利：授予用户的是原文所述的«a limited, revocable, non-exclusive and non-transferable license»，用于使用工具的内容；已读页面上免费档与付费档在商业使用上没有差别来源, 已核查 2026-08-07

Wordtune. 所输入文本的权利仍归用户，原文为：«You retain all right, title, and interest in and to your User Content»；同时，使用本服务的许可被授予 «solely for your own personal purposes»，而摘要产生的派生文本 «shall be used by the User for personal use only»。按资费有差异：所阅读的 Terms of Use 覆盖个人资费，而 «Our Business Terms govern use of Wordtune Business» 是另一份未曾阅读的文件来源, 已核查 2026-08-07

未提到套餐 — 17

条款对所有人谈论权利的方式相同，或者在所读到的内容中没有发现差别。

Anyword. 条款写道：«the Customer shall have all right and title in the Generated Content and Company shall not claim ownership with respect thereto»；同时工具保留使用材料的权利«on an aggregate and anonymized basis, to improve the Anyword Platform» 来源, 已核查 2026-08-04

Colossyan. 在付费并遵守规则的前提下允许商业使用：«Subject to the full payment of fees and compliance with the requirements of the Acceptable Use Policy, upon creation of a piece of Customer Created Content… Colossyan grants Customer a perpetual, limited, revocable, non-exclusive, worldwide, royalty-free license to use Colossyan… 来源, 已核查 2026-08-07

Gling. 条款没有谈剪辑成品的权利和成果的商业使用：它们是按网站条款写的，给出的是«a non-transferable, non-exclusive, revocable, limited license to use and access the Site solely for your own personal, noncommercial use»（原文），而«User Content»一节只限于用户责任和可接受使用规则。对一款付费剪辑工具来说，«个人非商业»的表述出人意料，但它指的是网站，而不是成果来源, 已核查 2026-08-04

Happy Scribe. 权利归用户所有，并明确延伸到工具的工作成果。原文：«you retain all rights, title, and interest in and to Your Content. This includes any content generated by the Services — such as transcripts, captions, translations, meeting recordings, or the notes and summaries produced by the Meeting… 来源, 已核查 2026-08-07

LTX Studio. 条款（Lightricks Ltd.）：«you own the Output (excluding any Stock Content or other third party materials therein)»，但紧接着有说明«you further acknowledge that pursuant to applicable law, it may be the case that you have no ownership rights, including copyrights, in the AI-generated Output»；商业使用与套餐挂钩：«you may… 来源, 已核查 2026-08-04

Luma (Dream Machine / Luma Agents). 条款：«Customer owns and retains all right, title, and interest in and to the Output and Luma hereby assigns to Customer all of Luma’s right, title, and interest in and to the Output»，但附有说明«Customer acknowledges and agrees that it can only use the Outputs for commercial purposes if the Outputs were… 来源, 已核查 2026-08-04

Maestra. 条款没有直接谈创作成果的权利。文中写道：«As between the parties, all Content, except for Customer Data (as defined below), is owned by Maestra or its licensors»，而对上传内容，用户保证«you own all rights, title, and interest in and to the Customer Data, or have otherwise secured all necessary… 来源, 已核查 2026-08-04

Notta. 上传内容的权利仍归用户。原文：«Customer retains all ownership rights in Customer Data»，工具方获得的许可为«to access, process, copy, export, and display Customer Data, only as reasonably necessary (a) to provide, maintain and update the Services»，以下为封闭清单。条款中没有单独规定转写和翻译成果本身的权利以及成果的商业使用来源, 已核查 2026-08-07

Rask AI. 条款：«You are permitted to use the Content you create through our Service outside of the Service, including in commercial contexts, provided that such use complies with all applicable laws, regulations, and these Terms»、«You retain full ownership rights to any Content created through our… 来源, 已核查 2026-08-04

Rytr. 条款写道：«You retain any and all of Your rights to any Content You create, post, or display on or through the Service»，同时用户授予工具许可«to use, modify, publicly perform, publicly display, reproduce, and distribute such Content on and through the Service»；条款中没有单独允许或禁止对所创作内容的商业使用来源, 已核查 2026-08-04

Sapling. 按合同而异，且该工具明确这样表述。限制的原文为：«unless you have signed a service agreement with Sapling stating otherwise, you shall not license, sell, rent, lease, transfer, assign, distribute, host, or otherwise commercially exploit the Service, whether in whole or in part, or any content displayed within the… 来源, 已核查 2026-08-07

Speechify. 关于对所创作内容的权利，条款只字未提；而且对工具本身的许可是这样给出的：«Speechify grants you… a limited, non-exclusive, nontransferable, revocable license to stream, download and make personal non-commercial use of the Services for which you have subscribed»；条款中没有单独允许对所生成配音的商业使用来源, 已核查 2026-08-04

Sudowrite. 条款只提到用户上传的内容：«All rights in and to Your Content remain with you»；关于对 AI 所创作内容的权利，以及对创作结果的商业使用或发布，条款未作说明来源, 已核查 2026-08-04

Synthesia. 所有权已明确：«Customer will own all Customer Data, but excluding the components of the Services and Synthesia Content» 来源, 已核查 2026-08-07

Udio. 本合集中最严格的限制，且对所有资费一视同仁。服务条款（12.11.2025 版）第 1.2 条原文：«You may not download copies of any Output or any other output generated by the Services»，另有义务不得复制、传播、展示、发布和分发输出 «on any personal device or on any streaming platform or user generated content platform (e.g., YouTube, Facebook, Instagram, TikTok… 来源, 已核查 2026-08-07

WellSaid Labs. 在价格页上，商业权利与套餐挂钩：Trial 为«3 download minutes per month (no commercial rights)»，Starter、Pro 和 Business 为«Full commercial rights»，通用副标题为«All with unlimited generation and commercial rights built-in»。而 https://www.wellsaid.io/terms-of-use 上的条款中没有这些内容：那里只规定了网站规则，许可是«to access and use the Website for your… 来源, 已核查 2026-08-04

Wisecut. 条款：«Except as provided in this Section 8, we obtain no rights under this Agreement from you (or your licensors) to Your Content»，即上传内容和剪辑成果的权利仍归用户；条款中没有单独允许或禁止商业使用；用户保证«you or your licensors own all right, title, and interest in and to Your Content» 来源, 已核查 2026-08-04

模型是否会拿你上传的内容做训练

会拿你的内容做训练 — 11

服务明确表示，它会把你输入的内容和生成的内容用于训练自己的模型。

Auphonic. 是，声明会以用户录音进行训练。服务条款：«your Content may be viewed and/or listened to by an Auphonic employee or any person contracted by Auphonic to work on our audio processing algorithms»（原文）；为改进算法，个别音频片段可以无限期保存，用户无法访问，也没有确定的期限。功能页面直接描述了持续训练：«Based on over five years of training with audio files from our web service, the algorithm… 来源, 已核查 2026-08-07

HeyGen. 对输入和产出都直接用于训练：用户授予许可 «to access, use, host, cache, store, reproduce, transmit, display, publish, distribute, and modify Your Content to operate, improve, promote and provide the Services and to develop new services and products, including to train or otherwise improve or modify our artificial intelligence and machine… 来源, 已核查 2026-08-07

HyperWrite. 是，原文为：本服务可将数据用于 «to improve the Software, including to train its machine learning algorithms applicable to its general user base and systems, and related products»；其中单独点名了信件内容和与服务的交互特征来源, 已核查 2026-08-07

Luma (Dream Machine / Luma Agents). 条款直接允许在输入和创作结果上训练：付费使用时，所授许可也包括为了«create, test, improve, train, or otherwise develop the artificial intelligence or machine learning models, systems, architecture, weights, or related technology used by Luma»，并且«will be perpetual and irrevocable with respect to any Input contained, incorporated, embodied, or otherwise… 来源, 已核查 2026-08-04

Murf AI. 条款中没有直接表明会以用户文本和音频训练产品模型。只有关于质量控制的保留：Murf 员工 «may use your Cloning Voice to create ... utterances, solely for internal quality assurance purposes»（原文），并且该服务可以试听音频样本 «to test the quality»。同时禁止用户把 Murf 的声音用于训练任何 AI 模型来源, 已核查 2026-08-07

quso.ai（原名vidyo.ai）. 是的，上传的视频会进入 AI 模型。隐私政策：«To enhance these features, we share anonymised data with these AI models. This include the video you uploaded to our app»（原文）。具体是哪些模型供应商 — 未点名来源, 已核查 2026-08-07

Sapling. 是，条款原文为：用户授予 Sapling 使用其内容的许可，«solely for the purposes of providing the Service and to improve the Service and develop new products and services»；隐私政策予以确认：«We collect text entered or otherwise transmitted by you» 以及 «Your content is stored by us and used only to improve the Service» 来源, 已核查 2026-08-07

Suno. 用输入内容训练已写进条款：用户保证不需要任何额外许可«arising out of or related to our use of your Submissions, including to create your Output or your Voice Model and/or to train, develop, fine-tune or otherwise improve the Service and any related artificial intelligence or machine learning models»；隐私政策予以确认：«We use certain User Activity… 来源, 已核查 2026-08-04

Udio. 是，明确声明。服务条款第 1.1 条原文：«You acknowledge and agree that we may use your Input Content and any Output to train or otherwise improve our Services, including our artificial intelligence and machine learnings model(s) related to the Services»（保留来源的拼写）。Input Content 指提示词、指令、歌词和上传的音频文件。用户还另外保证自己拥有上传内容的全部权利，包括将其用于 «to create… 来源, 已核查 2026-08-07

Wisecut. 条款直接要求同意用于训练：«You consent to our use of Your Content and any associated data to train and improve our artificial intelligence models. In such circumstances, we will not publicly post or share any of Your Content»；条款没有描述拒绝这种使用的方式来源, 已核查 2026-08-04

Wordtune. 是，对 User Content 的许可包含将其用于 «administer and make improvements to the Software and Services (including, to improve the algorithms underlying the Software and the Services)» 的权利 —— 原文如此；该许可被称为永久且不可撤销来源, 已核查 2026-08-07

不拿来训练 — 8

服务明确否认这一点。同样归入这里的还有只有在你同意的情况下才可能训练的情形：未经同意就不训练，对读者来说这就是«不会»的答案。

Amberscript. 否，隐私政策中明确否认。原文：«The audio and transcripts that are being generated, will never be used for training purposes of the language models that are used»。另外规定，订购人工校对时，文件和草稿转写对在编的转写员可见：原文为«your audio or video file and the automatically generated transcript will be visible to our transcribers and to our operations team in… 来源, 已核查 2026-08-07

Cleanvoice AI. 没有，该服务本身明确禁止训练。Terms & Conditions 第 11.7 条：所保存的内容不用于 «for training artificial intelligence models or for any purpose unrelated to the provision of the Services»（原文）来源, 已核查 2026-08-07

Colossyan. 三者中最直接的否认：«Colossyan will not use… Customer Materials to train any AI model or algorithm, with the sole exception that… Colossyan may use Recordings to train the AI model that creates and operates Customer’s Custom Created Avatar. Such AI model is exclusively used by Colossyan to provide the Custom Created Avatar for… 来源, 已核查 2026-08-07

Happy Scribe. 默认为否：只有在用户明确同意的情况下才可能用于训练。原文：«If you have not opted in, Your Content uploaded on or after the Effective Date is not used to train or maintain any machine-learning models»。开启同意后，上传内容将用于«develop, train and improve machine-learning models»（原文）来源, 已核查 2026-08-07

LALAL.AI. 没有。隐私政策原文：«the OmniSale GMBH does not assume the rights to any uploaded audio and video data, does not use user files for artificial intelligence training or other content»。在 30.07.2026 版的服务条款中未触及训练这一主题来源, 已核查 2026-08-07

LTX Studio. 条款授予 Lightricks«worldwide, non-exclusive, fully-paid, royalty-free, irrevocable, sublicensable and transferable license to use the Input you provide, the Output and your Account information including your or any third-party’s name, voice, image, likeness»，其中包括用于«for Lightricks’ internal purposes to improve our… 来源, 已核查 2026-08-04

ProWritingAid. 否，原文为：«When using ProWritingAid, your texts will not be used to improve the quality of our services»；同处还说明，语法检查完成后 «all submitted texts and their improvements will be deleted»，而算法是用专职专业文字编辑校订过的文本训练的来源, 已核查 2026-08-07

Sonix. 否，隐私政策中明确否认。原文：«Sonix does not use your content or Customer Data to train our machine learning or generative AI models»。另外规定，通过 OAuth 接入第三方 AI 应用时，工具只传递«only the data necessary to fulfill the requests you initiate through that application»，此后适用对方的政策来源, 已核查 2026-08-07

会训练，但有保留条件 — 2

同一份文件里两者兼有：训练是有的，但受到限制。把它归结为«是»或«否»，等于在两个方向上都说了谎。

Recraft. 条款：«You acknowledge and agree that Recraft may use your Inputs you provide and the Assets produced by the Services to train, fine-tune, and otherwise improve Recraft’s artificial intelligence models and systems. This training use is enabled by default, but you may opt out at any time by adjusting your preferences in… 来源, 已核查 2026-08-04

Synthesia. 存在基于使用数据的改进，但有明确保留：改进 «will not result in… (d) the improvement or fine tuning of any artificial intelligence components unless integrated with the Services being used by Customer; or (e) the creation or development of any artificial intelligence models that may be provisioned independent of the Services»（原文照录）。另外单独约定了… 来源, 已核查 2026-08-07

答复无法归结为一个词 —— 6

该表述既不能明确读作«是»，也不能明确读作«否»。它不作解释——逐字原样印出。

Anyword. 条款授予工具«a non-exclusive, perpetual, worldwide right and license to use, copy and reproduce the Content solely to (i) provide the Customer the Services; and (ii) provided that the Content is first aggregated and anonymized»——也就是说，上传的内容在汇总和去标识化之后用于改进平台；文本中未描述拒绝这一条件的方式来源, 已核查 2026-08-04

Bria. 关于自有模型，工具原文写道«Every asset in our training pipeline comes from commercial agreements with 30+ data partners»和«No scraping. No synthetic data. No gray areas»；关于用户上传内容，只说到在微调时原文为«our architecture keeps your data isolated and auditable»。已读页面上没有关于用户上传内容和提示词进入通用模型训练的直接表述来源, 已核查 2026-08-07

Let's Enhance. 隐私政策：工具对上传内容应用自动化系统，原文为«for security purposes, fraud prevention, compliance with legal and regulatory requirements, investigations of potential misconduct, product development and improvement, research, and customer or technical support»；没有关于用上传内容训练模型的表述来源, 已核查 2026-08-07

Playground. 条款授予工具对提示词、上传图像和结果的永久许可：原文为«a perpetual, worldwide, non-exclusive, sublicensable, no-charge, royalty-free, irrevocable copyright license to use, copy, reproduce, process, adapt, modify, publish, transmit, prepare Derivative Works of, publicly display, publicly perform, sublicense, and/or distribute text prompts and… 来源, 已核查 2026-08-07

Resemble AI. 条款允许基于衍生数据进行内部训练：Content 的使用目的是 «for the purposes of having Resemble AI synthesizing and generating an AI voice model»（原文），而 Aggregated Data «may be used for ... internal learning and training»（原文），并可用于调试、修复错误和产品开发。文中没有直接表明原始用户录音会进入产品模型的训练来源, 已核查 2026-08-07

SOUNDRAW. 条款没有谈用户输入内容的训练问题，也没有描述拒绝方式。该工具回答的是另一个问题——它自己训练数据的来源：«SOUNDRAW is an ethical AI music generator, trained using input from our in-house music team. Since we never use stolen material for training, our tracks will always be safe from copyright strikes»（原文）。这是对相邻问题的回答，而不是对所问问题的回答来源, 已核查 2026-08-04

能否拒绝

Amberscript. 没有描述单独的拒绝机制，也不需要：对所有用户默认否认用上传录音训练来源, 已核查 2026-08-07

Happy Scribe. 是：同意是自愿给出的，并可随时撤回。原文：«withdraw your opt-in at any time from your account settings or by emailing dataprotection@happyscribe.com» 来源, 已核查 2026-08-07

HyperWrite. 是，原文为：«you shall have the right to opt-out of the general systems»，拒绝之后数据只用于针对该具体用户训练的模型；申请通过邮件发送至 info@othersideai.com；同处还提醒，拒绝后用户将无法从基于通用系统训练的模型中获益来源, 已核查 2026-08-07

Krea. 按资费不同，这是主要发现。通用的 Terms of Use（2024 年 5 月 20 日版）完全没有谈用用户数据训练模型，尽管授予 Krea«worldwide, non-exclusive, perpetual, royalty-free, fully paid, sublicensable and transferable license to use, edit, modify, truncate, aggregate, reproduce... and otherwise fully exploit such User Submissions»；其中没有为… 来源, 已核查 2026-08-07

LTX Studio. 存在拒绝途径，但只是部分的，且仅针对个人数据。Terms of Service §6.2 授予 Lightricks 对 Input、Output 及账户数据（包括 «name, voice, image, likeness»）的许可，用途之一为 «(iii) for Lightricks' internal purposes to improve our products and services and for training purposes of any artificial intelligence-based models, now known or existing in… 来源, 已核查 2026-08-07

quso.ai（原名vidyo.ai）. 同意以上传时勾选复选框的方式作出：«By checking this box, you allow us to share anonymised data to enhance app features and functionality»（原文）。也就是说，不勾选即可拒绝。政策未描述撤回已作出同意的单独程序来源, 已核查 2026-08-07

Recraft. 所有用户都可以拒绝，开关在个人资料设置中。Terms of Service 的«Licenses to Recraft»一节原文：«You acknowledge and agree that Recraft may use your Inputs you provide and the Assets produced by the Services to train, fine-tune, and otherwise improve Recraft's artificial intelligence models and systems. This training use is… 来源, 已核查 2026-08-07

Sonix. 无需单独拒绝，因为默认就否认用客户数据训练。撤回机制只针对第三方接入：原文可以«revoke a connector's access at any time through your Sonix account settings or by contacting us» 来源, 已核查 2026-08-07

Sudowrite. 无可拒绝：该工具声明根本不用用户文本训练模型。FAQ 中对«Does Sudowrite use my writing to train models?»的回答原文为：«We are constantly listening to writers' feedback to improve Sudowrite. However, we do not use your writing to train AI models»。Terms and Conditions 并不与之矛盾：它们只给出许可«to host, store, process, display, and transmit Your… 来源, 已核查 2026-08-07

WellSaid Labs. 无可拒绝之事：服务协议明确禁止用客户数据训练 —— 原文为 «For clarity WellSaid will not use Customer Data to train its or any third party's AI models»（Customer Data 一节）；Customer Data 被定义为 Input 与 Output；关于用户自建的声音模型另有说明，WellSaid «will not share with other customers»；另外约定，使用数据（Usage Data，不含 Customer… 来源, 已核查 2026-08-07

服务是否给 AI 的作品加标注

平台要求对 AI 生成的内容加标注；为标注负责的是发布者。各工具自己对此的说法见下文。

2short.ai. 该服务在任何资费上都不承诺给成果加水印：功能区块中写着 «Say goodbye to watermarks and hello to 1080p quality»（原文），而据该服务称，各资费之间仅在分析小时数上有差别。免费资费没有单独的规则。元数据、C2PA 及其他 AI 来源标记均未被提及来源, 已核查 2026-08-07

Auphonic. 没有 AI 标注：服务条款和定价页中没有提到表示机器处理的水印、元数据或 C2PA。资费差别确实存在，但属于另一类 —— 声音插入：对比表中 «Includes Auphonic jingle» 一行对 Free 是 Yes，对 Monthly and One-Time 以及 Yearly and Business 是 No。这是品牌片头音，而不是使用了 AI 的标记来源, 已核查 2026-08-07

Bria. 工具在结果中嵌入署名：条款原文禁止«remove, obscure, or alter any attribution, watermarks, or metadata embedded in Outputs by Company's attribution technology»；此条未说明套餐之间的差别，也未按名称提到 C2PA 来源, 已核查 2026-08-07

Checksub. 免费方式下视频导出带该工具的水印（对比表中的«With Watermark»），所有付费资费无水印（«No Watermark»）。这是品牌水印，而不是说明字幕或译制由 AI 完成的标注；可访问页面上没有把成果标注为 AI 内容的要求来源, 已核查 2026-08-07

Colossyan. 免费资费有水印，去除水印从 Professional 资费起出售（定价页面）。条款中对水印、元数据和 C2PA 只字未提；同时其中明确禁止把 AI 创作的内容冒充为人类创作来源, 已核查 2026-08-07

Dubverse. 已读条款和价格页中没有把成果标注为 AI 内容的要求，也没有元数据或 C2PA。相反的规定是有的：两档付费资费都明确声明«No Water Mark»，即导出视频不会带水印来源, 已核查 2026-08-07

Fliki. 可见水印因资费而异：Free 为«Includes Fliki watermark»（原文），Standard 和 Premium 在对比表中为«No watermark»。关于元数据、C2PA 以及任何其他隐藏的 AI 来源标记，定价页面和条款都没有提及来源, 已核查 2026-08-07

Happy Scribe. 使用条款和价格页都没有声明把成果标注为 AI 内容。另有一点不同：免费资费的视频导出带有该工具的水印——«Export files in DOCX, TXT, SRT and MP4 (watermarked)»（原文），付费资费提供«Remove watermark from video exports»。这是品牌水印，而不是 AI 标注来源, 已核查 2026-08-07

HeyGen. 免费档有水印，去水印由付费档出售（价格页）。而条款本身对水印、元数据和 C2PA 只字未提，审核规则只要求为 18+ 内容加标：«clearly labeled at the outset of the video as mature content» 来源, 已核查 2026-08-07

LALAL.AI. 标注义务落在用户身上。«AI Transparency, Synthetic Content, and Regulatory Compliance» 一节援引（欧盟）2024/1689 号条例（EU AI Act）：生成功能的输出 «may legally constitute deepfakes or synthetic content»，接着原文写道：«You are solely responsible for clearly and distinguishably disclosing to the end audience that the audio content has… 来源, 已核查 2026-08-07

Let's Enhance. 条款和价格页均未提及 AI 生成内容的标注（元数据、C2PA）；免费档的水印被声明为该档的限制，而不是 AI 标记来源, 已核查 2026-08-07

Luma (Dream Machine / Luma Agents). 条款：«Luma reserves the right, in its sole discretion and without prior notice, to embed watermarks, content credentials, provenance metadata, or other identifying information in or associated with Output, including as may be required by applicable Laws»，禁止删除或规避它们；标注的义务落在用户身上：«If you use any Output… 来源, 已核查 2026-08-04

Mubert. 有水印，但只在下载之前。原文：«Mubert watermark is added to prevent fraudulent activity. The watermark will be deleted once you download the track to your device» —— 也就是说下载后的文件中不留标记。没有把成果标注为 AI 生成的要求。资费之间的差别属于另一类：免费的 Ambassador 不加标记，但必须做文字署名 —— 附 Mubert Render 链接和 #mubert 话题标签，付费资费则无需署名来源, 已核查 2026-08-07

Pika. 该服务在任何资费上都不加可见水印：«Download videos with no watermark»（原文）一行出现在全部四档资费上，包括免费资费。这一点各资费并无差别。关于元数据、C2PA 及其他隐藏的 AI 来源标记，定价页只字未提来源, 已核查 2026-08-07

Pixelcut. 已读页面均未提及 AI 生成标记（水印、元数据、C2PA）；同时价格页直接承诺免费导出不带水印来源, 已核查 2026-08-07

quso.ai（原名vidyo.ai）. 关于水印，该服务只从反面提及，并且只针对付费资费：Lite 为«Share crisp, watermark-free videos»，移动版 Plus 为«No watermarks on exported videos»（原文）。免费资费的说明中既未提到有水印，也未提到没有。元数据、C2PA 及其他 AI 来源标记在任何地方都未被提及来源, 已核查 2026-08-07

Recraft. 条款中有«Transparency and Marking of AI-Generated Content»一节：工具«may (i) embed machine-readable metadata, watermarks, or other provenance information into Assets generated by the Services; (ii) make available or link to a tool that enables users and third parties to detect whether content was generated… 来源, 已核查 2026-08-04

Resemble AI. PerTh 水印是 Resemble 的自有技术；对开源模型而言，标注被声明为强制且自动的：«Every clip is watermarked with PerTh at generation, before it leaves the model»（原文，Chatterbox Multilingual）；对 DramaBox 则是 «Every output watermarked with PerTh at generation»。该标记被声明为不可听且能抵抗转码。在服务条款中，Resemble Watermarker… 来源, 已核查 2026-08-07

Spikes Studio. 可见水印明确按资费区分，该服务对此直言不讳。FAQ：«Do the videos have watermarks? Yes, our free plan includes a watermark. To remove it, you'll need to subscribe to any of our paid plans»（原文）。免费资费卡片中重复标注 «Has watermark»，PRO+ 和 Enterprise 则是 «No watermark»。关于元数据和 C2PA 只字未提来源, 已核查 2026-08-07

Synthesia. 可接受使用规则禁止去除标记：禁止 «Removing, deactivating or disabling any «watermarks» or other mechanisms of the Services that are designed to help validate provenance or differentiate between human-generated and AI-generated content»（原文照录）。至于服务默认加上的究竟是什么，文件没有说明来源, 已核查 2026-08-07

如何使用

要查的是套餐，而不只是服务。 对成果的权利，更多取决于你用的是哪个套餐，而不是哪个工具。2. 文件的沉默不是许可。 如果没有提到权利、训练或拒绝，那就是没有答案；往对自己有利的方向解释，和往不利的方向解释一样危险。3. 给 AI 作品加标注的要求来自平台，而不是来自工具。 请看平台要求什么：为标注负责的将是发布内容的人。

关于方法。 每一行都是一条事实，并附服务文档的链接。答复的方向（«训练» / «不训练» / «有保留»）由一条带六个对照示例的规则加以归类；规则未能归类的内容，放入单独一节并逐字原样印出，而不是归入«是»或«否»。

另请参阅：用 AI 创作内容 · 平台对 AI 内容有什么要求 · 入门门槛