阿里云对象存储 OSS
提供方：阿里云计算有限公司
用途：提供英语语音包清单和压缩包下载。
触发方式：你在设置页对某个词库和口音主动点击“下载”后触发，不会在首次启动时静默下载。
可能处理的信息：所选词库与口音对应的资源路径、IP 地址、请求时间，以及完成内容分发和安全防护所必需的网络日志；不包含昵称、学习进度、错词、星级、WPM 或连续打卡记录。
控制方式：不点击下载即可避免该请求；已下载语音包可在设置中删除。
隐私规则：阿里云隐私政策
第三方服务与权限清单
本清单按当前 iOS、Android 与 HarmonyOS 发布版本的实际代码和数据流整理。除下列可选语音服务外，本应用不接入广告、用户行为分析、社交分享、支付或账号登录 SDK。第三方服务或处理方式变化时，我们会同步更新清单，并在依法需要时另行征得同意。
一、联网服务
设备系统文本转语音服务
提供方：iOS 上为 Apple Inc. 及适用的关联服务主体；Android 上为设备厂商或用户当前启用的系统语音引擎提供方；HarmonyOS 上为华为终端有限公司及相应服务主体。
用途：在未命中本地语音包时朗读英语单词，查询或下载设备离线音色；设备支持且必要时可能回退至在线语音。
触发方式：同意隐私政策后进入启用自动朗读的练习，或主动点击发音按钮。首次启动同意前不会触发。
可能处理的信息：待朗读的单词文本、语言或口音参数，以及提供下载或在线语音所必需的设备网络信息。具体处理以设备当前启用的系统语音引擎规则为准。
控制方式：可在设置中关闭“单词自动朗读”，并不再点击发音；下载本地语音包后，本应用会优先播放本地音频。
服务说明：iOS 用户可在系统设置的辅助功能与语音相关设置中管理系统音色；Android 用户可在系统设置中查看当前启用的文本转语音引擎；HarmonyOS 服务说明见华为 Speech Kit 文档。
二、设备权限
网络访问能力
声明：iOS 网络连接不需要单独的运行时权限；Android：android.permission.INTERNET；HarmonyOS：ohos.permission.INTERNET。
目的：下载你主动选择的语音包，以及允许系统语音服务按设备能力下载离线音色或提供在线朗读。
是否可关闭：该权限通常不提供运行时开关。你可通过不下载语音包、关闭自动朗读并不点击发音来避免对应联网场景。
触觉反馈
声明：HarmonyOS：ohos.permission.VIBRATE；iOS 与 Android 版本使用系统触觉反馈且不声明敏感权限。
目的：提供键盘敲击和答题反馈。
是否可关闭：可在应用设置中关闭振动。
当前版本不申请相机、麦克风、位置、通讯录、日历、电话、短信、相册或文件读取权限。
三、本地开源组件说明
本应用使用 Flutter、Riverpod、Drift/sqflite、audioplayers、Dio、archive 和 crypto 等开源组件完成界面、状态管理、本地数据库、本地音频、下载、解压和完整性校验。这些组件在本应用中的用途不包含广告或行为分析，也不会因为被引入就自动向其开源维护者发送学习数据。实际联网数据流仅包括上方列明的语音功能。
Third-Party Services & Permissions
This list reflects the current iOS, Android, and HarmonyOS release code and data flows. Other than the optional voice services below, the App includes no advertising, behavioral analytics, social-sharing, payment, or account-login SDK. We will update this list if providers or processing changes and obtain additional consent when required.
1. Network services
Alibaba Cloud Object Storage (OSS)
Provider: Alibaba Cloud Computing Co., Ltd.
Purpose: Deliver English voice-pack manifests and archives.
Trigger: Only after you select Download for a dictionary/accent in Settings; it is not silently downloaded on first launch.
Possible data: Selected dictionary/accent resource path, IP address, request time, and network/security logs necessary for delivery. No nickname, progress, wrong words, stars, WPM, or streak is included.
Control: Do not select Download; downloaded packs can be removed in Settings.
Privacy terms: Alibaba Cloud Privacy Policy
Device system text-to-speech service
Provider: On iOS, Apple Inc. and the applicable service entity; on Android, the device manufacturer or the system speech engine currently enabled by the user; on HarmonyOS, Huawei Device Co., Ltd. and the applicable service entity.
Purpose: Pronounce English words when local audio is unavailable, query/download offline voices, and, where supported and necessary, provide online speech.
Trigger: After consent, when a practice session has automatic pronunciation enabled or when you tap pronunciation. It is not triggered before first-launch consent.
Possible data: Word text, language/accent parameters, and device/network information necessary for voice download or online speech, subject to the rules of the system speech engine currently enabled on the device.
Control: Disable automatic pronunciation and avoid tapping pronunciation. The App prefers local audio after a voice pack is installed.
Documentation: Huawei Speech Kit documentation
2. Device permissions
Network access
iOS network connections require no separate runtime permission; Android declares android.permission.INTERNET; HarmonyOS declares ohos.permission.INTERNET. This is used for a voice-pack download you initiate and for system speech services to download offline voices or provide online speech as supported by the device. Avoid the corresponding network use by not downloading packs and by disabling/avoiding pronunciation.
Haptic feedback
HarmonyOS declares ohos.permission.VIBRATE. iOS and Android use system haptic feedback and declare no sensitive permission. Haptic feedback can be disabled in App Settings.
The current release does not request camera, microphone, location, contacts, calendar, phone, SMS, photo-library, or file-read permission.
3. Local open-source components
The App uses Flutter, Riverpod, Drift/sqflite, audioplayers, Dio, archive, and crypto for UI, state, local database, local audio, downloading, extraction, and integrity checks. Their inclusion does not automatically transmit learning data to open-source maintainers and they are not used for advertising or behavioral analytics. Actual network data flows are limited to the voice features listed above.