ImageCat Privacy Policy

Welcome to ImageCat ("we" or the "App"). We highly value your privacy and data security. We fully recognize the importance of user privacy and personal information protection. We will take appropriate measures as required by law to protect the security and controllability of your personal information. Based on these principles, we have formulated this Privacy Policy.

This Privacy Policy applies to all products and services provided by ImageCat. When you use our products or services, we may collect and use your personal information. This Policy explains how we collect, use, manage, share, transfer, disclose, protect, and store your personal information, as well as how you may access, correct, supplement, delete, or copy such information.

When using certain features, we may collect sensitive information with your consent. Refusing to provide such information may prevent you from using specific features, but it will not affect your ability to use other general functions of the App.

ImageCat uses third-party services to provide the functions described below. Our backend forwards selected user content to Volcengine for AI processing, and images are stored with Alibaba Cloud OSS. Routing a request through our own API does not mean the content is processed only by ImageCat. We disclose the recipients, data, purposes, and retention practices in this Policy.

ImageCat accesses only the device capabilities needed for a feature you actively use. Voice input requires microphone permission. Images are provided only when you select them through the system interface or otherwise choose to submit, save, or share them. These capabilities are not activated without your action or the authorization required by the operating system.

Table of Contents

1. Definitions
2. How We Collect and Use Your Personal Information
3. Local Storage and Similar Technologies
4. How We Share, Transfer, and Publicly Disclose Your Personal Information
5. How We Protect and Store Your Personal Information
6. Your Rights
7. Protection of Minors' Information
8. Notices and Revisions
9. How to Contact Us

1. Definitions

(1) Personal Information

Personal information means all kinds of information related to an identified or identifiable natural person, recorded electronically or otherwise, excluding anonymized information.

Information covered by this Policy includes: a CloudKit user-record identifier or random installation identifier used for identity verification and stored by ImageCat in hashed form; your ImageCat account ID and optional nickname; IP addresses, request identifiers, security and operation logs, quota and usage records; App Store subscription and entitlement status; text you submit, relevant conversation messages, design requirements and plans, images you select as inputs or references, generated images, and related generation records; and audio and transcripts produced when you actively use voice input. Images, text, or audio may contain personal or sensitive information depending on what you submit.

(2) Sensitive Personal Information

Sensitive personal information refers to information that, once leaked or illegally used, may easily lead to harm to personal dignity or personal/property safety, including biometrics, religious beliefs, specific identities, medical health, financial accounts, travel traces, and personal information of minors under the age of 14.

(3) ImageCat

"ImageCat" under this Policy refers to the ImageCat iOS app, the official ImageCat website, and any other related products or services operated by the ImageCat developer.

(4) Affiliates

Affiliates refer to any entity that controls, is controlled by, or is under common control with a party. "Control" means direct or indirect ownership of more than 50% of equity, voting rights, or management rights.

(5) Anonymization

Anonymization means processing personal information so that a specific natural person cannot be identified and the information cannot be restored.

(6) De-identification

De-identification means processing personal information so that a specific natural person cannot be identified without additional information.

(7) Virtual Goods

Virtual goods refer to digital products or services that can be used online within ImageCat, including membership services and future digital service offerings.

2. How We Collect and Use Your Personal Information

Before collecting and using your personal information, we will inform you of the purpose, the categories of information involved, how such information will be used, and the relevant retention scope, and we will obtain your consent where required.

We use information for the specific functions, security, troubleshooting, and service-operation purposes described in this Policy. Aggregated statistics may be used to evaluate and improve the App. De-identification is not the same as anonymization: data that can still be linked to you remains subject to personal-information safeguards. Using identifiable user content for a separate purpose, such as a model-training program, requires a separate disclosure and the authorization required by law; permission to process an AI request does not itself authorize such a program.

Where sensitive personal information is involved, or where personal information would be publicly displayed or used for purposes beyond those previously agreed, we will obtain your prior consent through an explicit action such as a checkbox or confirmation.

You may revoke device permissions in system settings, delete relevant content, or contact us at hello@imagecat.app to withdraw permission for processing. Revoking microphone access stops new recording; it does not itself delete information already transmitted. Account cancellation and deletion requests are handled subject to the retention rules below and applicable law.

We may compile statistics on the use of our products or services and share such statistics with the public or third parties to demonstrate general usage trends, provided that no personally identifiable information is included.

ImageCat uses CloudKit to establish a passwordless identity when iCloud is available. The App sends a CloudKit user-record identifier and a short-lived CloudKit web-authentication token to the ImageCat backend for verification; the backend stores a one-way hash of the verified identifier rather than the original identifier. This process is not Sign in with Apple. We do not receive your Apple ID password, Apple login name, or email address. If iCloud is unavailable or restricted, we use a random installation identifier to create a device-based guest identity and store only a one-way hash on the backend. If CloudKit later becomes available, the guest identity may be linked to the verified CloudKit identity.

We use account, request, subscription, quota, and generation information to establish and secure your ImageCat identity, provide AI conversation and image generation or editing, restore membership status, prevent abuse, troubleshoot failures, and maintain service records. Input and reference images are uploaded to Alibaba Cloud OSS for processing; generated images are also stored there for delivery and history. Our database records prompts, input and output image references, account and session associations, generation status, time, and usage information. Images and related records may be accessed as necessary to deliver and maintain the service, investigate failures, and evaluate generation quality and improve the App; these purposes do not constitute permission for a separate model-training program. When you delete a local history item, the App also requests deletion of the related image objects; local deletion does not itself erase all database records, backups, or third-party records.

AI conversation and image processing

For AI conversation, our backend sends your submitted text, relevant conversation messages, design requirements and plans, image labels and version context, and selected images needed for the current request to Volcengine's Doubao service through the Volcano Engine Ark API. Images are made accessible using temporary signed URLs. For image generation or editing, we send the generation prompt, relevant input or reference images, and output parameters such as image size or aspect ratio to the same provider. This is third-party processing of user content even when the provider receives the content through an API and returns only a reply or image. The conversation API does not write a complete chat archive to our business database. Complete conversation history is saved locally as described in Section 3; submitted generation prompts and relevant records are retained on the backend.

When you actively start voice input, the App records audio and streams it through our backend to Volcengine's Doubao speech-recognition service, which converts it to text. Audio is sent during recording, not only after you finish. The recognition request also includes your internal ImageCat account ID as the provider's user identifier, audio-format parameters, and request identifiers. If live recognition fails, the App may upload a temporary recording to our backend for file recognition; a temporary access URL lets the provider read it. This fallback audio is held in a short-lived backend cache, deleted after the recognition attempt, and given a 150-second expiry as a failsafe. The current recognition flow does not archive recordings in OSS or the generation database. The App removes its temporary recording during cleanup. Transcripts are returned to your input field and may enter local drafts or conversations, subsequent AI requests, and generation prompts. Where transcript segments are submitted with an image-generation request, they may also be saved with the generation record. Successful speech usage may record a hash of the transcript for quota accounting and duplicate-request checks.

3. Local Storage and Similar Technologies

The App stores settings and credentials locally, including language, appearance, AI-badge preferences, cached membership status, creation history, your ImageCat account ID, and an access token protected by the device keychain or equivalent system storage. It also saves conversation messages and design requirements in local files to restore your work, and unsent text drafts in device storage. Unsent drafts have a 30-day retention limit, with expired drafts removed when draft storage is next accessed or updated. These local records are distinct from the content transmitted for cloud AI processing. Deleting the related history or account clears applicable local records through the App's deletion flow.

The current ImageCat public website does not use advertising cookies. Hosting and network providers may still process ordinary technical logs, such as IP address, request time, browser type, and requested URL, to deliver and secure the website.

4. How We Share, Transfer, and Publicly Disclose Your Personal Information

(1) Sharing

• We may share personal information after obtaining your explicit consent.
• We may share information when required by law, regulation, or compulsory governmental request.
• We may share necessary information with our affiliates for the purposes stated in this Policy, subject to the same protection requirements.
• We may exchange information with other companies or organizations where necessary to comply with law, enforce agreements, or protect the rights, property, or safety of ImageCat, users, or the public, such as for fraud prevention.
• When CloudKit identity is available, Apple processes the CloudKit account-status, user-record, and web-authentication requests needed to verify the passwordless identity. This is not Sign in with Apple, and ImageCat does not receive your Apple login name or email address through this process.
• App Store and RevenueCat process purchase, receipt, subscription, restoration, and entitlement information needed to provide and synchronize PRO membership.
• Volcengine (北京火山引擎科技有限公司): provides Doubao conversation and image services through Volcano Engine Ark, and Doubao speech recognition. It receives the text, relevant conversation and design context, participating images or temporary image-access URLs, audio or temporary audio-access URLs, and technical identifiers and parameters described in Section 2, according to the function used. See the Volcengine Privacy Policy and the applicable Ark service terms.
• Alibaba Cloud (阿里云计算有限公司): provides OSS image storage. It receives input, reference, and generated image files, object paths and account-related storage identifiers, and the technical information necessary for storage, access control, delivery, and deletion. The current speech-recognition flow does not store recordings in OSS. See the Alibaba Cloud Privacy Policy.

Permission and third-party safeguards

Sharing personal data with a third-party AI service requires clear disclosure of the data, recipient, and purpose, and your explicit permission before transmission. This Policy, a link to it, continued use of the App, and system permission to access the microphone or selected photos do not by themselves constitute that permission. If you decline, the corresponding cloud AI processing must not start. You can contact us at hello@imagecat.app to withdraw permission or request deletion, and stop submitting content to the corresponding feature. Withdrawal does not undo completed processing or override lawful retention obligations.

We require every third party receiving user data to provide protection at least equal to that stated in this Policy and required by applicable App Store rules, including purpose limitation, confidentiality, appropriate security, necessary retention limits, and cooperation with deletion and rights requests. Provider privacy policies and applicable service terms describe their obligations; merely linking to them does not replace our responsibility for selecting and overseeing these services. We use these APIs for inference and transcription; these API calls are not dataset uploads or model-fine-tuning jobs.

Provider data use and protection

Volcengine's Product and Service Agreement and applicable Data Protection Agreement describe restrictions on customer-data use, security obligations, and deletion or anonymization. Alibaba Cloud's OSS service terms provide that customer business data belongs to the customer, restrict unauthorized use and disclosure, require confidentiality, and provide for deletion according to customer instructions, subject to applicable legal requirements and separately agreed terms. We require these safeguards to cover the user data handled for ImageCat and to meet the protection standard stated above.

For Doubao conversation and image generation through Ark, the published Doubao Model Service Agreement provides that inputs and outputs are not used to train, retrain, or improve foundation models without the customer's separate agreement. Accepting a separate data-use authorization or participating in a collaboration reward program is an exception. Permission for an ImageCat creation request does not authorize those additional purposes. Before arranging an additional use of your personal data for model training or optimization, we will explain the data, recipient, and purpose and obtain your separate explicit permission.

Under the applicable Ark and Doubao terms, Volcengine may review submitted and generated content for legal compliance and content safety, including maintaining risk filters and unlawful-content feature libraries. Necessary content and records may also be processed or retained for troubleshooting that we request and for investigating abnormal calls or security alerts. These activities are subject to the applicable service rules and necessary retention limits. They are distinct from ImageCat's temporary audio cache and do not mean that the provider deletes every request immediately after returning a result.

(2) Transfer

• We will not transfer your personal information to any company, organization, or individual except where you have given explicit consent, where required by law, where consistent with relevant agreements, or where necessary in a merger, acquisition, or asset transfer. In such cases, we will require the successor entity to continue to be bound by this Policy or to seek your consent anew.

(3) Public Disclosure

• We will publicly disclose your personal information only when you explicitly request or consent to the disclosure, or where required by law, regulation, or compulsory administrative or judicial process.

(4) Exceptions

An exception to consent applies only where a specific applicable legal basis permits the particular disclosure. This is not a blanket exception for AI conversation, image generation or editing, or speech recognition; ordinary provision of these features does not replace explicit permission for third-party AI sharing.

5. How We Protect and Store Your Personal Information

(1) Protection

We use security measures that meet industry standards to protect the personal information you provide against unauthorized access, disclosure, use, modification, damage, or loss. We may use encryption technologies such as TLS/SSL, access control, de-identification, and other trusted protection mechanisms.

We may analyze the ImageCat account ID, identity mode, IP address, request identifiers, operation logs, quota events, and service errors to prevent abuse and protect service security. Input, reference, and generated images are stored only as needed to process requests, deliver results, preserve the history you choose to keep, and meet applicable security or legal requirements.

We maintain dedicated internal procedures and personnel access controls to protect personal information and ensure that only authorized staff may access relevant data.

We will retain your personal information only for the minimum period necessary to achieve the purposes stated in this Policy or as otherwise required by applicable law. If a personal information security incident occurs, we will notify you in accordance with law and, where necessary, report it to relevant authorities.

Retention differs by data type. Current speech-recognition audio is transient as described in Section 2; the 150-second fallback-cache expiry applies to our cache and is not a promise about the provider's retention. Local chat history remains available until removed through the applicable deletion flow. Generation prompts, image references, submitted transcript segments, and usage records are retained as needed for the purposes above and applicable legal obligations. Pending generation requests are stored for queueing and recovery; their request bodies are cleared when jobs finish, fail, or are cancelled. Ordinary service logs hide message and prompt bodies by default and retain operational information such as request identifiers, timing, counts, and errors. Diagnostic logging may include AI replies, which can contain submitted content; diagnostic information is used to investigate service failures. Service logs, database records, and backups are subject to necessary security, accounting, troubleshooting, and legal retention; contact us at hello@imagecat.app to request deletion or further information. Third-party processing and any security-related retention are governed by the applicable service terms and our data-protection requirements, not by our local cache expiry.

(2) Storage

Personal information collected and generated during operations in the People's Republic of China will be stored within the territory of China unless otherwise required by law or authorized by you. After account cancellation, we will delete or anonymize your personal information as required by law. If we terminate our services or operations, we will notify you in advance and delete or anonymize your information afterward as required.

6. Your Rights

• Access: you may review your information through the relevant pages in the ImageCat app.
• Correction or supplementation: if you discover that the information we process about you is inaccurate, you may request correction through the channels described in the app or by contacting customer support at hello@imagecat.app.
• Deletion: you may request deletion where our processing violates law, where we collected or used your information without consent, where we violated our agreement with you, where you no longer use our services, or where we no longer provide products or services.
• Account cancellation: you may cancel the current ImageCat identity through Account Management in the App. The App clears local creation history and establishes a new identity for later use. The previous identity is marked as cancelled, and related personal information will be deleted or anonymized according to the applicable retention period, except where otherwise required by law.

We may be unable to respond to certain requests in circumstances permitted by law, including matters involving national security, public interests, criminal investigations, abuse of rights, serious harm to others' legitimate interests, or trade secrets.

7. Protection of Minors' Personal Information

We attach great importance to the protection of minors' personal information. ImageCat is primarily intended for adults. If you are under 18, you should obtain the consent of your guardian before using our products or services. If you are the guardian of a minor under 14, you should read and agree to this Policy on behalf of your ward before they use our products or services.

If we discover that we collected personal information of a minor without verifiable prior consent from a parent or legal guardian where such consent was required, we will endeavor to delete the relevant data promptly.

8. Notices and Revisions

To provide better services and adapt to changes in the law and our business, we may update this Privacy Policy from time to time. We will publish any changes on this page.

For material changes, we may also provide more prominent notice such as in-app messages, website announcements, or other appropriate means.

Material changes may include, for example, major changes to our service model, the purposes or methods of processing personal information, changes in ownership due to business restructuring, or significant changes to your rights and how they may be exercised.

Changes to a third-party AI recipient, the categories of data sent, the processing purpose, or material protection and retention practices will be disclosed before the changed processing starts. Where the existing authorization does not cover the change, new explicit permission is required before transmission. Continued use alone does not authorize a new AI recipient or processing purpose. A model-version change within the same provider does not by itself change the recipient; any accompanying change in data handling remains subject to these requirements.

9. Contact Us

ImageCat 隐私协议

欢迎使用ImageCat（以下简称"我们"或"本应用"）。我们非常重视你的隐私与数据安全，深知用户个人隐私及信息的重要性，并非常重视用户（以下简称"你"）的隐私和个人信息保护。我们将按照法律法规的要求，采取相应措施保护你个人信息的安全可控。基于前述理念和目的，我们制定《ImageCat隐私协议》。

本隐私协议适用于ImageCat提供的所有产品和服务。你在使用我们的产品与/或服务时，我们可能会收集和使用你的个人信息。我们希望通过本隐私协议向你说明我们如何收集、使用、管理、共享、转让、披露、保护和保存你的个人信息，以及我们为你提供的访问、更正、补充、删除、复制这些信息的方式。

当你使用一些功能时，我们会在获得你的同意后，收集你的一些敏感信息。拒绝提供这些信息仅会使你无法使用相关特定功能，但不影响你正常使用我们产品或服务的其他功能。

ImageCat 使用下述第三方服务实现相应功能：我们的后端会将参与处理的用户内容发送至火山引擎进行 AI 处理，图片会保存于阿里云 OSS。请求先经过我们自己的 API，并不代表内容仅由 ImageCat 处理。本协议说明数据接收方、数据范围、处理目的及保存方式。

ImageCat 只会访问你主动使用的功能所必需的设备能力。语音输入需要麦克风权限；只有当你通过系统界面选择图片，或主动提交、保存、分享图片时，我们才会处理相应内容。未经你主动操作或操作系统要求的授权，这些能力不会被启用。

目录

1. 定义
2. 我们如何收集和使用你的个人信息
3. 本地存储和同类技术
4. 我们如何共享、转让、公开披露你的个人信息
5. 我们如何保护和保存你的个人信息
6. 你的权利
7. 未成年人信息的保护
8. 通知和修订
9. 如何联系我们

一、定义

（一）个人信息

个人信息是以电子或者其他方式记录的与已识别或者可识别的自然人有关的各种信息，不包括匿名化处理后的信息。

本协议涉及的信息包括：用于身份验证并由 ImageCat 以哈希形式保存的 CloudKit 用户记录标识或随机安装标识；ImageCat 账户 ID 和可选昵称；IP 地址、请求标识、安全及操作日志、额度和使用记录；App Store 订阅和会员权益状态；你提交的文字、相关对话消息、设计需求和方案、你选择的输入图或参考图、生成结果图及生成记录；以及你主动使用语音输入时产生的音频和转写文字。图片、文字或音频是否包含个人信息或敏感信息，取决于你提交的具体内容。

（二）敏感个人信息

敏感个人信息是一旦泄露或者非法使用，容易导致自然人的人格尊严受到侵害或者人身、财产安全受到危害的个人信息，包括生物识别、宗教信仰、特定身份、医疗健康、金融账户、行踪轨迹等信息，以及不满十四周岁未成年人的个人信息。

（三）ImageCat

本隐私协议内所称 ImageCat，是指由 ImageCat 开发者运营的 ImageCat iOS App、ImageCat 官方网站以及其他相关产品或服务。

（四）关联方

关联方是指控制某一方的、或被某一方所控制的、或与某一方共同受控制于同一实体的任何企业。控制是指直接或间接拥有该企业 50% 以上的股权、投票权或管理权。

（五）匿名化

匿名化是指个人信息经过处理无法识别特定自然人且不能复原的过程。

（六）去标识化

去标识化是指个人信息经过处理，使其在不借助额外信息的情况下无法识别特定自然人的过程。

（七）虚拟商品

虚拟商品是指ImageCat内可以通过在线形式使用的数字产品与/或服务，包括但不限于会员服务及未来可能上线的数字产品与/或服务。

二、我们如何收集和使用你的个人信息

在收集和使用你的个人信息前，我们会事先告知你收集目的、信息类型、使用方式等，并在需要时取得你的同意。

我们按照本协议说明的具体功能、安全、故障排查及服务运营目的处理信息，并可使用汇总统计评估和改进 App。去标识化不等于匿名化：仍可与你关联的数据继续按照个人信息要求保护。将可识别的用户内容用于独立目的，例如模型训练计划，需要另行告知并取得依法所需的授权；处理一次 AI 请求的许可本身不包含这类独立用途。

在收集你的敏感个人信息、公开展示你的个人信息或将你的个人信息用于你非事先同意的其他用途时，我们会通过勾选等显著方式事先征求你的同意。

你可以在系统设置中撤回设备权限、删除相关内容，或通过 hello@imagecat.app 联系我们撤回处理许可。撤回麦克风权限可停止新的录音，但不会自动删除已经发送的信息。注销及删除请求按照下述保存规则和适用法律处理。

我们可能会对产品与/或服务使用情况进行统计，并与公众或第三方共享这些统计信息，以展示整体使用趋势，但该等统计信息不包含你的任何身份识别信息。

当 iCloud 可用时，ImageCat 使用 CloudKit 建立免密码身份。App 会将 CloudKit 用户记录标识和短时效 CloudKit Web 验证令牌发送给 ImageCat 后端进行验证；后端保存的是验证后标识的单向哈希，而不是原始标识。这个过程不是“通过 Apple 登录”，我们不会获取你的 Apple ID 密码、Apple 登录姓名或邮箱。如果 iCloud 不可用或受限，我们使用随机安装标识建立设备访客身份，并仅在服务端保存其单向哈希。CloudKit 后续恢复可用时，访客身份可能与经验证的 CloudKit 身份关联。

我们使用账户、请求、订阅、额度和生成信息来建立并保障 ImageCat 身份，提供 AI 对话及图片生成或编辑、恢复会员状态、防止滥用、排查故障并维护服务记录。输入图和参考图会上传至阿里云 OSS 供处理，生成结果图也保存在该服务中供交付和历史访问。数据库记录提示词、输入及输出图片的关联、账户和会话关联、生成状态、时间及使用信息。我们可在提供和维护服务、排查故障、评估生成质量及改进 App 所需的范围内访问图片和相关记录；这些用途不构成独立模型训练计划的授权。当你删除本地历史项目时，App 也会请求删除相关云端图片对象；本地删除本身不会同步清除所有数据库记录、备份或第三方记录。

AI 对话及图片处理

使用 AI 对话时，我们的后端会通过火山方舟 API 将你提交的文字、相关对话消息、设计需求和方案、图片标签及版本上下文，以及当前请求所需的相关图片发送至火山引擎的豆包服务。图片通过临时签名地址供服务读取。使用图片生成或编辑时，我们会向同一服务商发送生图提示词、参与处理的输入图或参考图，以及图片尺寸、比例等输出参数。即使对方通过 API 接收内容并仅返回回复或图片，这仍属于第三方对用户内容的处理。对话接口不会将完整聊天档案写入我们的业务数据库；完整对话历史按照第三节说明保存在设备本地，提交的生图提示词及相关记录则会保存在后端。

当你主动启动语音输入时，App 会录音，并经我们的后端实时发送至火山引擎的豆包语音识别服务转换为文字。音频在录音过程中就会发送，并非仅在录音结束后发送。识别请求还会以你的 ImageCat 内部账户 ID 作为服务方的用户标识，并包含音频格式参数和请求标识。实时识别失败时，App 可将临时录音上传至我们的后端进行文件识别，并通过临时访问地址供服务方读取。该兜底音频保存在后端短时缓存中，识别尝试结束后删除，另设 150 秒过期时间作为兜底；当前识别流程不会将录音归档至 OSS 或生成数据库。App 会在清理流程中删除本地临时录音。转写文字返回输入框后，可进入本地草稿或对话、后续 AI 请求及生图提示词。若转写片段随生图请求提交，也可能随生成记录保存。成功的语音使用记录可保存转写文字的哈希摘要，用于额度计费和重复请求校验。

三、本地存储和同类技术

App 会在本地保存语言、外观、AI 角标偏好、缓存会员状态、创作历史、ImageCat 账户 ID，以及受设备钥匙串或等效系统存储保护的访问令牌。App 还会将对话消息和设计需求保存为本地文件以恢复创作，并在设备存储中保存未发送的文字草稿。未发送草稿按 30 天保留上限处理，超期草稿在后续访问或更新草稿存储时清理。这些本地记录与为云端 AI 处理而发送的内容不同。删除相关历史或账户时，App 会通过相应删除流程清理适用的本地记录。

当前 ImageCat 公开网站不使用广告 Cookie。网站托管和网络服务商仍可能为了提供和保护网站而处理常规技术日志，例如 IP 地址、请求时间、浏览器类型和请求 URL。

四、我们如何共享、转让、公开披露你的个人信息

（一）共享

• 在获取明确同意的情况下共享；
• 根据法律法规规定或政府主管部门强制性要求对外共享；
• 与我们的关联公司共享必要的个人信息，并受本隐私协议约束；
• 为遵守法律、执行协议、保护ImageCat、你或其他用户的权利、财产或安全等目的而与其他公司和组织交换信息。
• 当 CloudKit 身份可用时，Apple 会处理验证免密码身份所必需的 CloudKit 账户状态、用户记录和 Web 验证请求。这不是“通过 Apple 登录”，ImageCat 不会因此获取你的 Apple 登录姓名或邮箱。
• App Store 和 RevenueCat 会处理提供并同步 PRO 会员所必需的购买、凭证、订阅、恢复购买和权益状态信息。
• 火山引擎（北京火山引擎科技有限公司）：通过火山方舟提供豆包对话及图片服务，并提供豆包语音识别。根据你使用的功能，该服务方接收第二节所述文字、相关对话及设计上下文、参与处理的图片或临时图片访问地址、音频或临时音频访问地址，以及相关技术标识和参数。请参阅火山引擎隐私政策及适用的火山方舟服务条款。
• 阿里云（阿里云计算有限公司）：提供 OSS 图片存储，接收输入图、参考图和生成结果图文件、对象路径及与账户相关的存储标识，以及存储、访问控制、交付和删除所需的技术信息。当前语音识别流程不会将录音保存在 OSS 中。请参阅阿里云隐私政策。

授权与第三方保护要求

向第三方 AI 服务发送个人数据，需要在发送前明确告知数据、接收方及用途，并取得你的明确许可。本协议、协议链接、继续使用 App，以及系统对麦克风或所选照片的访问许可，本身均不替代这项明确许可。如你拒绝，相应云端 AI 处理不得开始。你可通过 hello@imagecat.app 联系我们撤回许可或请求删除，并停止向相应功能提交内容。撤回不会逆转已经完成的处理，也不免除依法保留记录的义务。

我们要求每一家接收用户数据的第三方提供不低于本协议及适用 App Store 规则要求的数据保护，包括限定处理目的、保密、适当的安全措施、必要的保存期限，以及配合删除和权利请求。服务商隐私政策和适用服务条款说明其义务，提供链接本身不替代我们选择及监督这些服务的责任。我们使用上述 API 进行推理及转写，这些 API 调用并非上传训练数据集或执行模型微调任务。

服务商的数据用途与保护

火山引擎的产品和服务协议及适用的数据保护协议说明客户数据的使用限制、安全义务以及删除或匿名化规则。阿里云的OSS 服务条款明确业务数据归客户所有，限制未经授权的使用和披露，约定保密义务，并规定按客户指令删除数据；适用法律要求及另行约定的情形除外。我们要求这些保护措施覆盖服务商为 ImageCat 处理的用户数据，并达到上述数据保护标准。

对于通过火山方舟提供的豆包对话及生图服务，公开的豆包大模型服务协议规定，未经客户单独同意，不将输入和输出用于基础模型的训练、重新训练或改进；另行接受数据使用授权或参加协作奖励计划属于例外。你对 ImageCat 一次创作请求的许可，不包含这些额外用途。我们如安排将你的个人数据用于额外的模型训练或优化，会在开始前说明数据、接收方及用途，并取得你的单独明确许可。

根据适用的方舟及豆包条款，火山引擎可能为履行法律要求和保障内容安全，对提交内容及生成结果进行审核，包括维护风险过滤机制和违法内容特征库。为处理我们主动申请的故障排查，以及调查异常调用或安全告警，服务商也可能处理或保留必要的内容及记录。这些处理受相应服务规则和必要保留期限约束，与 ImageCat 的临时录音缓存不同，并不代表服务商在返回结果后立即删除每一次请求的全部信息。

（二）转让

• 我们不会将你的个人信息转让给任何公司、组织和个人，但事先获得你明确同意、法律法规要求、符合与你签署的相关协议约定，或在合并、收购、资产转让等情况下除外。届时我们会要求新的持有人继续受本隐私协议约束或重新征得你的授权同意。

（三）公开披露

• 我们仅在根据你的明确需求或在法律法规、强制性行政执法或司法要求必须提供的情况下，才会公开披露你的个人信息。

（四）例外情形

只有在具体适用法律依据允许相关披露的范围内，才可适用免于同意的例外。这不构成对 AI 对话、图片生成或编辑、语音识别的概括性豁免；正常提供这些功能不能替代第三方 AI 数据共享所需的明确许可。

五、我们如何保护和保存你的个人信息

（一）你个人信息的保护

我们已使用符合行业标准的安全防护措施保护你提供的个人信息，防止数据遭到未经授权访问、公开披露、使用、修改、损坏或丢失。我们会采取包括加密、访问控制、匿名化或去标识化处理等在内的一切合理可行的措施来保护你的个人信息。

我们可能分析 ImageCat 账户 ID、身份模式、IP 地址、请求标识、操作日志、额度事件和服务错误，以防止滥用并保护服务安全。输入图、参考图和生成结果图仅在处理请求、交付结果、保留你选择保留的历史记录，以及满足适用的安全或法律要求所必需的范围内存储。

我们建立了专门的管理制度、流程和组织，确保只有授权人员方可访问信息，并要求其承担保密义务。

我们只会在达成本隐私协议所述目的所需的期限内保留你的个人信息，除非法律法规另有要求。如发生个人信息安全事件，我们将按照法律法规的要求及时通知你，并向监管部门报告。

不同数据的保存方式不同。当前语音识别音频按照第二节说明临时处理；150 秒兜底缓存过期时间仅适用于我们的缓存，不代表第三方服务商的保存期限。本地聊天历史会保留至通过适用删除流程清理。生图提示词、图片关联、随生图提交的转写片段及使用记录，会在实现上述目的及履行适用法律义务所需的期限内保存。排队中的生图请求会保存用于调度和恢复，任务完成、失败或取消后清除其请求正文。普通服务日志默认隐藏消息及提示词正文，保留请求标识、耗时、数量和错误等运行信息。诊断日志可能包含 AI 回复，其中可能带有提交的内容；诊断信息用于排查服务故障。服务日志、数据库记录和备份按照安全、计费、故障排查及法律要求所需的期限保留；你可通过 hello@imagecat.app 联系我们请求删除或了解具体情况。第三方处理及其安全相关保留受适用服务条款和我们的数据保护要求约束，并不受我们本地缓存过期时间控制。

（二）你个人信息的保存

我们在中华人民共和国境内运营过程中收集和产生的你的个人信息将存储于境内，除非法律有明确规定或者获得你的明确授权。

在你注销账号后，我们将根据法律法规的要求对你的个人信息进行删除或匿名化处理。如果我们终止服务或运营，我们会提前通知，并在终止后按照法律规定删除或匿名化处理你的个人信息。

六、你的权利

• 访问：你有权在ImageCat App 的对应页面内查询你的信息。
• 更正、补充：当你发现我们处理的关于你的个人信息有错误时，你有权要求我们作出更正，也可以通过 hello@imagecat.app 联系客服提出补充申请。
• 删除：如果我们处理个人信息的行为违反法律法规、未征得你的同意、违反了与你的约定，或者你不再使用我们的产品或服务，你可以请求删除个人信息。
• 注销：你可以在 App 内的“账户管理”中注销当前 ImageCat 身份。App 会清除本地创作历史，并为后续使用建立新身份。原身份将被标记为已注销，相关个人信息将按适用的保留期限删除或匿名化，法律法规另有要求的除外。

在与国家安全、公共利益、刑事侦查、明显恶意或滥用权利、涉及商业秘密等情形相关时，我们可能无法响应你的请求。

七、未成年人的个人信息保护

我们非常重视对未成年人个人信息的保护。ImageCat产品主要面向成年人，若你是不满十八周岁的未成年人，在使用我们的产品与/或服务前，应事先取得你的监护人的同意。若你是未满十四周岁未成年人的监护人，在你的被监护人使用产品及服务前，你应为其阅读并同意本隐私协议。

如果我们发现自己在未事先获得可证实的父母或法定监护人同意的情况下收集了未成年人的个人信息，则会设法尽快删除相关数据。

八、通知和修订

为给你提供更好的服务以及随着ImageCat业务的发展，我们可能会根据服务更新的情况及法律法规的相关要求更新本隐私协议。我们会在本页面内公布对隐私协议作出的任何变更。

对于重大变更，我们还会提供更为显著的通知，例如 App 内消息、网站公告或其他适当方式。重大变更包括但不限于服务模式变化、处理目的或类型变化、所有者变更以及你参与个人信息处理方面的权利和行使方式发生重大变化等。

第三方 AI 接收方、发送的数据类别、处理目的或重要保护及保存方式发生变化时，应在变更后的处理开始前明确告知。原授权未覆盖变更的，应在发送前重新取得明确许可。继续使用本身不构成对新增 AI 接收方或处理目的的授权。同一服务商内部更换模型版本本身不改变数据接收方，但同时发生的数据处理变化仍应遵守上述要求。

九、联系我们