隐私政策Privacy Policy

更新日期：2026 年 10 月 31 日Updated: October 31, 2026

欢迎使用珩AI（以下简称「本服务」）。我们非常重视你的隐私与个人信息保护。本政策说明我们如何收集、使用、存储与保护你的信息。使用本服务即表示你已阅读并同意本政策。Welcome to Heng AI (the "Service"). We value your privacy and the protection of your personal information. This policy explains how we collect, use, store, and protect your information. By using the Service, you agree to this policy.

一、我们收集的信息1. Information We Collect

• 账号信息：当你使用游客登录时，我们会基于设备唯一标识生成账号；当你使用 Google 登录时，我们会收集你的邮箱、昵称与头像。Account information: When you sign in as a guest, we create an account based on your device's unique identifier. When you sign in with Google, we collect your email, name, and avatar.
• 你主动提供的内容：你在对话中发送的消息、上传的文件、以及设置的记忆、技能、MCP 配置等。Content you provide: messages you send, files you upload, and the memories, skills, and MCP configurations you set.
• 你自行接入的外部服务数据：你配置的 MCP 服务在你下达指令时返回的数据（仅用于执行该次指令）。Data from external services you connect: data returned by the MCP servers you configure when you issue an instruction (used only to carry out that instruction).
• 日志与运行数据：为保障服务稳定与安全，我们可能记录基本的访问日志。Log and operational data: we may record basic access logs to keep the Service stable and secure.

二、我们如何使用信息2. How We Use Information

• 提供、维护与改进本服务的核心功能；Provide, maintain, and improve the core features of the Service;
• 调用第三方 AI 模型以响应你的指令：文本模型 DeepSeek（deepseek-v4-flash，端点 api.deepseek.com）、视觉与向量模型阿里云通义千问（qwen3-vl-plus / text-embedding-v4，端点阿里云百炼）；以及联网搜索、执行你配置的 MCP；Call third-party AI models to respond to your requests: the text model DeepSeek (deepseek-v4-flash, endpoint api.deepseek.com) and the vision/embedding models Alibaba Qwen (qwen3-vl-plus / text-embedding-v4, endpoint Alibaba Bailian); plus web search and the MCP servers you configure;
• 数据出境与存储：上述 AI 推理仅用于当次请求、处理你主动选定的内容，不持久化用于训练，也不用于训练或改进任何模型；Data residency & retention: the AI inference above is used only for the current request, processing the content you explicitly select; it is not persisted for training and is never used to train or improve any model;
• 用于账号安全、风控与防止滥用；Ensure account security, risk control, and prevent abuse;
• 在你授权范围内，用于改进本服务面向你的功能与体验；不将你的数据用于广告或与面向你的功能无关的用途。Improve the user-facing features and experience of the Service within the scope of your authorization; never use your data for advertising or purposes unrelated to the user-facing features.

二之一、知识库资料的处理2a. How Knowledge Base Content Is Processed

当你把文档、图片或扫描件导入知识库时，这些内容会经过以下处理，仅用于让你的资料可被检索与问答：When you import documents, images or scans into the knowledge base, the content is processed as follows, solely to make your material searchable and answerable:

• 解析：文档解析与存储在本服务内完成，不会外发；Parsing: document parsing and storage happen inside the Service and are not sent out;
• 向量化：为支持语义检索，资料摘录与关键词会发送给阿里云百炼的 embedding 服务（text-embedding-v4）；Embedding: to support semantic retrieval, excerpts and keywords from your material are sent to Alibaba Bailian's embedding service (text-embedding-v4);
• 扫描件 OCR：图片型页面会发送给阿里云百炼的视觉模型（qwen-vl-ocr）做文字识别；Scanned-page OCR: image-based pages are sent to Alibaba Bailian's vision model (qwen-vl-ocr) for text recognition;
• 编译：长文档会被拆解并发送给 DeepSeek 生成知识卡；Compilation: long documents are split and sent to DeepSeek to generate knowledge cards;
• 上述外发内容均不用于训练。你可以在知识库的单库设置中关闭「云编译」与「向量嵌入」，此时该库内容完全不出内网（代价是失去语义检索与自动编译）。None of the above is used for training. You can disable cloud compilation and embedding per knowledge base, in which case that vault's content never leaves the Service (at the cost of semantic retrieval and automatic compilation).

三、信息的存储与保护3. Data Storage and Protection

• 我们采取加密存储、访问控制等技术与管理措施保护你的数据；We protect your data with technical and administrative measures such as encryption and access control;
• 你配置的 API Key 等敏感凭据均加密存储；Sensitive credentials you configure (such as API keys) are stored encrypted;
• 我们仅在实现本政策所述目的所必需的期限内保留你的信息。We retain your information only as long as necessary for the purposes described in this policy.

三之一、保留期限3a. Retention Periods

• 账号与内容：账号、资料、会话与记忆保留至你删除相关内容或注销账号；删除后主存储立即删除。Account and content: your account, files, conversations, and memories are retained until you delete them or close your account; once deleted, the primary storage is deleted immediately.
• 备份副本：不超过 30 天，随后按备份轮换周期自动清除。Backups: retained for no more than 30 days, then removed by the backup rotation cycle.
• 访问与运行日志：不超过 90 天，用于安全审计与故障排查。Access and operational logs: retained for no more than 90 days for security auditing and troubleshooting.

三之二、Cookie 与本地存储3b. Cookies and Local Storage

我们使用必要的 Cookie 与浏览器本地存储来维持登录会话、记住语言与界面偏好、进行安全校验；不使用广告或跨站追踪 Cookie。你可以在浏览器设置中清除这些数据，但可能需要重新登录。We use essential cookies and browser local storage to maintain your sign-in session, remember language and interface preferences, and perform security checks. We do not use advertising or cross-site tracking cookies. You can clear this data in your browser settings, though you may need to sign in again.

三之三、跨境处理3c. International Processing

为提供服务，你的数据可能在你所在地区以外的司法辖区处理（例如 Google 登录由 Google 处理，AI 推理由相应的模型服务商处理）。我们会通过合同、访问控制与加密等措施保护你的数据。To provide the Service, your data may be processed in jurisdictions outside your region (for example, Google processes Google Sign-In, and AI inference is handled by the relevant model providers). We protect your data through contracts, access controls, encryption, and other safeguards.

四、与第三方的共享4. Sharing with Third Parties

• 为提供服务，我们可能将必要的数据传递给大模型服务商、搜索服务商等第三方；传递给第三方的仅为完成你的指令所必需的内容；To provide the Service, we may share necessary data with third parties such as AI model providers and search providers; only the content necessary to carry out your instructions is shared;
• 我们仅在必要范围内共享，并要求第三方遵守相应的保密与安全义务；We share only to the necessary extent and require third parties to comply with confidentiality and security obligations;
• 除法律要求或取得你的明确同意外，我们不会出售你的个人信息。We do not sell your personal information, except as required by law or with your explicit consent.
• 主要处理方包括 DeepSeek（文本模型）、阿里云百炼（视觉与向量模型）、Google（账号登录）以及运行本服务所依赖的云基础设施；具体以你实际启用的功能为准。Our main processors include DeepSeek (text models), Alibaba Cloud Bailian (vision and embedding models), Google (account sign-in), and the cloud infrastructure used to run the Service, depending on the features you enable.

五、Google 登录5. Signing in with Google

你可以使用 Google 账号登录本服务。该流程仅申请 openid、email、profile 三个基础作用域，用于获取你的用户唯一标识、邮箱、昵称与头像，以便创建账号、展示昵称头像与找回登录。我们不申请、也不访问任何受限或敏感作用域。你可以随时在 Google 账号的「安全性 → 第三方应用与权限」中撤销授权。You may sign in with a Google account. That flow requests only the three basic scopes openid, email and profile, used to obtain your unique user identifier, email, name and avatar so we can create your account and display your nickname/avatar. We request and access no restricted or sensitive scopes. You can revoke access at any time under your Google Account's "Security → Third-party apps & services".

六、Google API 数据使用（有限使用声明）6. Google API Data Use (Limited Use Disclosure)

本服务对通过 Google API 获取的信息的使用与向其他应用的传输，遵守 Google API 服务用户数据政策，包括其中的有限使用（Limited Use）要求。The Service's use and transfer to any other app of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.

• 我们获取的 Google 数据范围：仅为登录所必需的账号基础信息 —— 用户唯一标识（sub）、邮箱、昵称与头像。What we receive from Google: only the basic account information needed to sign you in — your unique identifier (sub), email, name and avatar.
• 我们不访问的数据：不申请、不读取 Gmail、Google Drive、Docs、Sheets、Slides、Calendar、Tasks、Contacts 等任何用户内容数据。What we do not access: we neither request nor read any user content from Gmail, Google Drive, Docs, Sheets, Slides, Calendar, Tasks or Contacts.
• 仅用于提供面向你的功能：上述信息仅用于创建与登录你的账号、在界面中显示昵称与头像。Only to provide user-facing features: the information above is used solely to create and sign you into your account and to display your nickname and avatar in the interface.
• 禁止用途：我们不会将 Google 用户数据用于投放广告、出售数据、构建广告画像、跨应用追踪，或用于与面向你的功能无关的其他目的（包括训练与相关功能无关的 AI 模型）。Prohibited uses: we do not use Google user data for advertising, selling data, building advertising profiles, cross-app tracking, or any purpose unrelated to the user-facing features (including training AI models unrelated to those features).
• 受限传输：我们不会将从 Google API 获取的信息传输给第三方，除非为提供上述登录功能所必需或为遵守法律所要求。Limited transfer: we do not transfer information received from Google APIs to third parties, except as necessary to provide the sign-in features described above or as required by law.

AI NCII 禁止：本服务不提供图像或视频生成能力，且不会使用 Google API 或本服务生成、编辑、传播 AI NCII（未经同意的 AI 生成亲密影像）。你也不得利用本服务制作或传播此类内容。AI NCII prohibition: the Service does not provide image or video generation, and does not use Google APIs or the Service to generate, edit, or distribute AI NCII (AI-generated non-consensual intimate imagery). You must not use the Service to create or distribute such content.

七、你的权利7. Your Rights

• 你可随时查看、更正或删除你的记忆、技能、会话与账号数据；You can view, correct, or delete your memories, skills, conversations, and account data at any time;
• 你可通过注销账号等方式要求删除个人信息；You can request deletion of your personal information, for example by deleting your account;
• 你可以在 App 内删除会话、资料或注销账号，也可以发邮件至 aplomb.chen@gmail.com 请求删除；我们会在合理期限内处理。You can delete conversations, files, or your account in the app, or email aplomb.chen@gmail.com to request deletion; we will process requests within a reasonable period.
• 你可随时撤回第三方授权或停止使用本服务。You can revoke third-party authorization or stop using the Service at any time.

八、未成年人保护8. Protection of Minors

本服务不面向未满 14 周岁的未成年人。如你为未成年人，请在监护人指导下使用。The Service is not intended for minors under 14. If you are a minor, please use it under the guidance of a guardian.

九、政策更新9. Policy Updates

我们可能适时更新本政策，更新后将在本页面公布。重大变更我们会以显著方式通知你。We may update this policy from time to time, and updates will be published on this page. We will notify you of material changes in a prominent manner.

十、联系我们10. Contact Us

如对本政策或我们的数据处理有任何疑问，请通过邮箱 aplomb.chen@gmail.com 与我们联系。If you have any questions about this policy or our data practices, please contact us at aplomb.chen@gmail.com.

珩AI · 私人知识库 · 随时分享Heng AI · Private knowledge base · share anytime