书林Collect ideas. Grow your knowledge.
隐私政策 / Privacy Policy
书林处理提供账号登录、书签收藏、网页采集、整理与搜索所需的数据。正文、图片和必要上下文可能经我们自营的 AI 平台发送至 OpenAI。我们不出售用户数据，不将其用于个性化广告或信用评估。您可管理书签、关闭自动采集，并通过本页邮箱提出隐私或删除请求。
1. 适用范围与联系方式
本政策适用于 Clipgrove · 书林网站及 Chrome 扩展“Clipgrove — Web Clipper & Bookmark Manager”。本服务由 Clipgrove · 书林个人开发者运营；在本政策中，“我们”指书林服务运营者。
服务网址：https://clipgrove.app4j.com。隐私联系邮箱：eric9530737@gmail.com。有关个人信息、第三方处理或数据删除的请求，均可发送至该邮箱。
2. 我们处理的信息
- 账号及身份验证信息：邮箱、账号标识、登录时提交的密码及用于身份验证的信息。扩展在浏览器本机存储 API Token，以便获得授权后访问您的书签库。
- 书签与用户输入：收藏的网址、标题、收藏时间、分组、标签、备注，以及您主动提交或修改的其他书签信息。
- 网页与附件内容：采集或上传的网页正文、HTML 页面归档、图片、支持的音视频或其他附件，以及处理生成的文字材料、摘要和标签。实际范围取决于页面可访问内容、您的操作和功能设置。
- 本机设置与任务状态：主题、采集开关、待处理任务和相关状态等，保存在当前浏览器本机存储中。
- 访问与运维信息：服务端和网络基础设施处理请求时记录的来源 IP、请求地址与时间、浏览器请求信息、响应状态、用量及错误诊断信息。日志可能包含收藏网址或与故障有关的内容片段。
保存的网址及相关信息属于浏览活动相关数据。我们围绕您选择收藏的内容提供功能，不为广告画像而建立您的完整浏览历史。我们不主动记录与本功能无关的鼠标轨迹或键盘输入，不要求用户提供健康记录、金融账户或私人通讯来使用书签服务。用户主动收藏的网页本身仍可能含有敏感内容，请谨慎选择。
3. 处理时机、目的与权限
您通过工具栏、网页右键菜单或其他收藏入口保存网页时，扩展会读取完成收藏所需的网址、标题和可访问内容。开启页面内容采集时，还会处理正文及可取得的图片等资源，用于归档、搜索和整理。我们使用浏览器存储、当前标签页临时权限、脚本执行、右键菜单及定时任务权限实现这些功能；网站访问权限用于连接书林服务和经授权的采集任务。
公众号链接自动采集默认关闭。您主动开启并授予公众号网站访问权限后，扩展可为已收藏但缺少正文的文章运行后台任务，包括打开采集页面、读取内容并上传至书林；首次开启也可能处理已有待补全收藏。关闭开关会暂停相关自动任务。
我们使用上述信息提供账号登录、保存与管理书签、提取和归档内容、生成摘要和标签、提供搜索，并维护这些功能的安全与可靠性。网站会使用保持登录所需的 Cookie；扩展使用本机存储保存配置与授权信息。
4. AI 处理与第三方服务
书林的 AI 请求经过我们自营的平台，再发送至 OpenAI 执行摘要、标签或图片识别等处理。相关请求可包含收藏正文、图片、标题及完成任务所需的上下文。当前服务可在收藏内容就绪后自动执行摘要或标签任务，因此 AI 处理不一定每次都由单独点击触发；该处理也不是全部在您的设备上完成。
| 处理方 | 用途与涉及的信息 |
|---|---|
| 书林运营者及其自营服务 | 运行书签库、存储与搜索、内容处理和 AI 请求转发；处理账号、书签、附件、必要请求信息及诊断记录。 |
| 阿里云（Alibaba Cloud） | 提供网站公网入口、网络接入及相关日志的基础设施服务；在其承载范围内处理网络请求、IP 地址和访问／运行日志。业务应用、书签存储及运维备份还使用运营者管理的自营主机和受限运维环境，并非全部位于阿里云。 |
| OpenAI | 接收完成摘要、标签或图片识别所需的正文、图片、标题与上下文，并返回模型结果。数据处理及保留还受实际使用接口、账户设置和适用协议约束。参见 OpenAI API 数据控制说明。 |
OpenAI 官方说明，API 数据默认不用于训练模型，除非账户主动选择共享。OpenAI 可能保留滥用监控日志，部分接口也可能保存应用状态；本服务未向用户承诺 OpenAI 的零数据保留。我们不将自身收集的用户内容用于出售、广告画像或与书签功能无关的模型训练。
外部 AI 服务可能在您所在国家或地区以外处理相关内容；我们未向用户承诺限定在某一个国家处理所有 AI 请求。访问来源网页或下载其资源时，来源网站也会依其政策处理相应网络请求。
5. 数据用途与访问限制
我们对通过 Chrome 扩展及相关 API 获取的信息的使用遵守 Chrome Web Store User Data Policy，包括 Limited Use 要求。
- 仅为提供、维护、安全保障及改进已披露的书签收藏、采集和整理功能使用必要信息。
- 不出售用户数据，不将其提供给广告平台或数据经纪商，不用于个性化广告、信用评估或贷款。
- 仅在提供已披露功能所必需、履行法律义务、防范欺诈或安全威胁，或适用政策允许且满足相应条件的情形下传输信息。涉及合并、收购或资产出售的数据转移，将满足适用通知和事先同意要求。
- 限制人工访问用户内容，仅在获得针对具体数据的同意、必要安全调查、履行法律义务，或符合政策的匿名汇总内部运营等允许情形下访问。
6. 存储、安全与实际保留安排
账号、书签、附件及派生材料保存在书林服务端；扩展设置、API Token 和任务状态等保存在当前浏览器本机。正式服务使用 HTTPS 传输，并以账号权限控制访问。我们采取必要的保护措施，但不声称所有存储都已加密，也不保证任何系统能完全消除风险。
为持续提供您的书签库，账号与书签内容保留至您删除相应内容、申请删除账号，或我们不再需要为披露的目的保留。以下是本政策更新时核实的日志和备份安排：
| 类别 | 实际保留与清理安排 |
|---|---|
| 公网入口访问与错误日志 | 每日检查，非空时轮转，最多保留52份历史归档，另有当前日志文件。该规则按归档份数执行，并不是严格的52天自动删除承诺；实际覆盖时间受访问情况和轮转执行影响。 |
| 书林应用容器日志 | 当前未设置按天到期或按大小轮转。日志可能持续保留至人工清理或相应容器移除；重启服务本身不会保证删除日志。删除请求会单独核对涉及的记录及依法需要保留的部分。 |
| 手工运维备份 | 迁移、部署或账号维护时生成的数据库／文件快照，在受限服务器及运维环境保存。目前没有统一的自动到期清理周期，不会因在界面删除书签而立即从已有快照中消失。可通过联系邮箱申请核对和清理涉及个人数据的备份。 |
| 应用内账号备份 | 当前自动备份未启用。启用后按账号配置的频率和保留天数执行，当前默认保留设置为30天；此默认设置不适用于上述手工运维备份。您可通过可用的备份管理功能管理应用内备份。 |
| 自营 AI 平台与 OpenAI | 自营平台处理必要用量与错误诊断记录，独立于网站入口的52份日志轮转规则；删除请求会一并核查该平台记录。OpenAI 的监控日志、应用状态和其他副本按适用接口及账户的数据控制规则处理，不随书林界面删除操作即时同步删除。 |
对于保留期较长或没有固定自动到期的资料，我们会依据其必要性、存储维护和经核实的删除请求处理；法律或安全原因需要保留时，会说明适用限制。备份恢复时，我们会核对已受理的删除请求，避免将应删除的数据重新用于正常服务。
7. 您的选择及数据请求
您可查看、修改和删除书签，关闭自动采集、调整采集设置，并在 Chrome 中撤销网站权限或卸载扩展。卸载扩展不会自动删除已上传至服务端的账号和书签。
您可通过 eric9530737@gmail.com 申请访问、导出、更正、删除个人信息或注销账号，也可询问 AI 处理和日志／备份范围。我们可能要求核实请求者身份，但不会要求您通过普通邮件发送登录密码或 API Token。
收到经核实的删除请求后，我们会核对在线数据、必要日志、运维备份及自营 AI 平台记录，处理能够删除的资料，并说明法律保留、共享日志拆分或第三方控制所造成的限制及后续安排。本政策不将界面删除描述为所有历史副本即时擦除。
8. 政策更新
数据类型、用途、服务方或保留安排发生变化时，我们会更新本页及日期，并按需要更新商店声明、在产品中提供适当提示或取得所需同意。有关本政策的问题，请联系本页邮箱。
Privacy policy
This policy applies to the Clipgrove website and the Chrome extension “Clipgrove — Web Clipper & Bookmark Manager”. The service is operated by the individual developer of Clipgrove · 书林. Privacy contact: eric9530737@gmail.com.
1. Information we process
We process account email addresses and identifiers; credentials submitted for authentication; the extension's API token; saved URLs, titles, timestamps, groups, tags and notes; captured or uploaded page text, HTML archives, images and supported attachments; derived text, summaries and tags; local settings and capture-task state; and necessary request, usage and diagnostic logs. Logs can include IP addresses, request URLs and times, browser request information, response status and content fragments relevant to an error.
The service operates on content selected for bookmarking. We do not build a complete browsing history for advertising, or intentionally record unrelated mouse movements or keystrokes. The bookmarking service does not require health records, financial accounts or private communications, although a page you choose to save may itself contain sensitive information.
2. Collection, purposes and permissions
When you save a page, the extension reads the URL, title and accessible content needed for bookmarking. Content capture can include text, images and other available resources. Browser storage, temporary active-tab access, packaged scripts, context menus, alarms and optional host permissions support these functions. Login cookies and local token storage support authentication.
Automatic WeChat capture is disabled by default. Once enabled by you with the relevant site permission, it can open saved articles with missing content, capture accessible content and upload it to Clipgrove, including previously saved articles awaiting completion. Turning the feature off pauses these automatic tasks.
We use data to authenticate users, save and organize bookmarks, archive content, generate summaries and tags, provide search, and maintain the security and reliability of these features.
3. AI processing and recipients
Relevant text, images, titles and necessary context are sent through our self-operated AI platform to OpenAI for summaries, tags and image recognition. Summarization and tagging can run automatically after content becomes available. This processing is not performed entirely on your device.
Alibaba Cloud provides the website's public network entry and related infrastructure/log storage within its service scope. The application, bookmark storage and operational backups also use infrastructure and restricted maintenance environments managed by the operator; not all data is stored on Alibaba Cloud.
OpenAI receives content required for model processing. Its API data controls documentation explains that API data is not used for training by default unless the account opts in, and that monitoring logs and some application state may be retained. We do not promise Zero Data Retention. AI processing may occur outside your country; we do not promise one processing country for every request. Source websites also process requests under their own policies when pages or resources are accessed.
4. Limited Use and access
Our use of information obtained through the extension and related APIs complies with the Chrome Web Store User Data Policy, including Limited Use requirements. We use necessary data only for the disclosed bookmarking purpose and related operations. We do not sell user data, provide it to advertising platforms or data brokers, or use it for personalized advertising, creditworthiness or lending. We do not use collected content for unrelated model training.
Transfers are limited to necessary disclosed functionality, legal obligations, security purposes, or other permitted circumstances with applicable conditions satisfied. Transfers in a merger, acquisition or asset sale will meet applicable notice and prior-consent requirements. Human access is limited to permitted situations such as specific user consent, necessary security investigations, legal obligations or compliant anonymized aggregate internal operations.
5. Storage, security and retention
Account, bookmark, attachment and derived content are stored server-side. Extension settings, tokens and task state are stored in the browser. Production traffic uses HTTPS and account access controls. We do not claim that every storage layer is encrypted or that all risks can be eliminated.
- Account and library data: retained to provide your library until deletion, an account-deletion request, or until no longer necessary for the disclosed purposes.
- Public-entry access/error logs: checked daily and rotated when nonempty, with up to 52 historical archives plus current log files. This is a file-count rule, not a strict 52-day deletion guarantee.
- Application container logs: no time-based expiry or size-based rotation is currently configured. They may remain until manual cleanup or container removal; a restart does not ensure deletion.
- Manual operational backups: migration, deployment and maintenance snapshots currently have no uniform automatic expiry. Copies remain in restricted server/maintenance storage until reviewed or removed; deleting a bookmark does not immediately erase existing snapshots.
- In-app backups: automatic backups are currently disabled. If enabled, they follow account frequency and retention settings; the current default retention setting is 30 days. This setting does not apply to manual operational backups.
- AI platform and OpenAI records: necessary usage/error records on our platform are separate from the website's log rotation. They are included when reviewing a deletion request. OpenAI records follow applicable endpoint/account rules and are not immediately erased by a deletion in the Clipgrove interface.
We review data based on necessity, maintenance and verified deletion requests, subject to legal and security requirements. When restoring backups, we check previously accepted deletion requests to avoid returning data that should be deleted to normal service.
6. Your choices, requests and changes
You can view, edit or delete bookmarks, disable automatic capture, change capture settings, revoke Chrome site permissions or uninstall the extension. Uninstalling does not delete uploaded server-side data.
Contact eric9530737@gmail.com to request access, export, correction, deletion or account closure, or information about AI processing and retention. We may verify your identity; do not email your password or API token. We review live data, relevant logs, backups and self-operated AI platform records, act on data we can delete, and explain legal, technical or third-party limitations and next steps. Interface deletion is not a promise of immediate erasure of all historical copies.
We update this page and its date when practices change, and update store disclosures, provide in-product notices or obtain consent as required.