edit | blame | history | raw

App Store Connect Submission Prep

Build

  • App: 私语输入法 / PrivateVoice Dictation
  • Bundle ID: com.shanghai3168.privatevoicedictation
  • Version: 2.1.28
  • Build: 20260604.0049
  • Platform: macOS
  • Recommended category: Productivity
  • Secondary category: Utilities
  • Minimum macOS for App Store sandbox build: 13.4
  • Support URL: https://shanghai3168.github.io/privatevoice-dictation-support/support.html
  • Privacy Policy URL: https://shanghai3168.github.io/privatevoice-dictation-support/privacy.html

Current Local Readiness

  • Xcode: 26.3
  • Upload tool: xcrun altool 26.10.1
  • App Store sandbox entitlements file: privatevoice.src/build/darwin/entitlements-appstore.plist
  • Mac App Store package script: privatevoice.src/scripts/build-macappstore-pkg-macos.sh
  • Apple App Store app and installer signing identities are installed.
  • Mac App Store provisioning profile is installed for com.shanghai3168.privatevoicedictation.
  • Local upload blocker remaining: App Store Connect upload credentials or an active Transporter session.

Required Apple Account Items

  1. App Store Connect app record for bundle ID com.shanghai3168.privatevoicedictation.
  2. Mac App Store provisioning profile for com.shanghai3168.privatevoicedictation.
  3. App signing certificate installed in Keychain:
  • Apple Distribution: <Team> (<TeamID>), or
  • legacy 3rd Party Mac Developer Application: <Team> (<TeamID>).
  1. Installer signing certificate installed in Keychain:
  • 3rd Party Mac Developer Installer: <Team> (<TeamID>).
  1. App Store Connect role with build upload permission.

Build Upload Package

Current signed upload package:

/Users/ar/Projects/PrivateVoice/Release/PrivateVoice-Dictation-2.1.28-build20260604.0049/PrivateVoice-Dictation-2.1.28-build20260604.0049-universal-macappstore.pkg
/Users/ar/Desktop/PrivateVoice-26.0604上架/02-上传包/PrivateVoice-Dictation-2.1.28-build20260604.0049-universal-macappstore.pkg

The package contains universal arm64 and x86_64 binaries for the app executable, libsherpa-onnx-c-api.dylib, and libonnxruntime.1.24.4.dylib.

Validate before upload:

pkgutil --check-signature "/Users/ar/Projects/PrivateVoice/Release/PrivateVoice-Dictation-2.1.28-build20260604.0049/PrivateVoice-Dictation-2.1.28-build20260604.0049-universal-macappstore.pkg"
xcrun altool --validate-app \
  -f "/Users/ar/Projects/PrivateVoice/Release/PrivateVoice-Dictation-2.1.28-build20260604.0049/PrivateVoice-Dictation-2.1.28-build20260604.0049-universal-macappstore.pkg" \
  -t macos \
  -u "<apple-id>" \
  -p "<app-specific-password>"

Upload:

xcrun altool --upload-app \
  -f "/Users/ar/Projects/PrivateVoice/Release/PrivateVoice-Dictation-2.1.28-build20260604.0049/PrivateVoice-Dictation-2.1.28-build20260604.0049-universal-macappstore.pkg" \
  -t macos \
  -u "<apple-id>" \
  -p "<app-specific-password>"

Transporter can also upload the same signed .pkg.

Chinese Metadata Draft

Name

私语输入法

Subtitle

私密离线语音输入

Promotional Text

离线语音识别、自定义词库和隐私保护，适合日常写作、办公记录和中英文输入。

Description

私语输入法是一款面向 macOS 的私密离线语音输入工具。按下快捷键开始说话，识别结果会自动输入到当前文本框，适合写邮件、写文档、聊天、做会议记录和处理日常办公文本。

语音识别在本地设备上完成，支持 SenseVoice、Qwen3-ASR、Moonshine English、Parakeet 等本地模型。应用不会收集、上传或存储你的语音、文本、词库或使用数据。自定义词库和识别历史仅保存在本机，由你自己管理。

主要功能：

  • 离线语音识别，数据不出设备
  • 中文、英文和中英混输场景支持
  • 自定义词库，修正常见误识别词
  • 可选择轻量默认模型或更高质量模型
  • 全局快捷键输入，适合在任意文本框使用
  • 本地识别历史，便于回看和整理

本应用基于开源项目 VoiceSnap 二次开发，遵循 MIT License。完整版权声明、许可文本、隐私政策和支持方式可在支持网站查看。

Keywords

语音输入,听写,离线识别,输入法,语音转文字,隐私,词库,办公,写作,中文输入

What's New

补全关于页中的苹果审核信息，明确说明本地离线识别、数据不出设备、开源项目归属、版权许可、隐私政策和支持反馈入口。

English Metadata Draft

Name

PrivateVoice Dictation

Subtitle

Private offline voice dictation

Promotional Text

Offline speech recognition, custom dictionaries, and privacy-first dictation for daily writing and productivity.

Description

PrivateVoice Dictation is a private offline voice input tool for macOS. Press a hotkey, speak, and the recognized text is typed into the active text field. It is designed for writing, email, chat, notes, meeting summaries, and everyday productivity.

Speech recognition is completed locally on your device with local models including SenseVoice, Qwen3-ASR, Moonshine English, and Parakeet. The app does not collect, upload, or store your voice, text, dictionary, or usage data. Custom dictionaries and recognition history stay on your Mac and remain under your control.

Key features:

  • Offline speech recognition with data staying on device
  • Chinese, English, and mixed Chinese-English dictation workflows
  • Custom dictionary for correcting commonly misrecognized words
  • Lightweight default models and optional higher-quality models
  • Global hotkey input for use in other apps
  • Local recognition history for review and organization

This app is a secondary development based on the open-source VoiceSnap project and follows the MIT License. Full copyright notices, license text, privacy policy, and support links are available on the support website.

Keywords

dictation,voice input,speech to text,offline,privacy,dictionary,macOS,writing,productivity

What's New

Updated the About page with App Review information covering local offline recognition, on-device data handling, open-source attribution, license details, privacy policy, and support links.

Privacy Nutrition Label Recommendation

  • Data Collection: Data Not Collected.
  • Tracking: No.
  • Third-party advertising: No.
  • Analytics: No.
  • Voice data: processed locally only; not collected by the developer.
  • Text/user dictionary/history: stored locally on device only; not collected by the developer.
  • Network use: model package downloads and opening support/privacy links. No voice/text/user data is uploaded.

App Review Notes

PrivateVoice Dictation is a macOS offline dictation app.

All speech recognition runs locally on the user's Mac. The app does not collect, upload, or store voice recordings, recognized text, user dictionary entries, recognition history, or usage analytics on any developer server.

Network access is used only to download optional local speech model packages and to open the support/privacy/license pages in the user's browser. Recognition remains local after models are installed.

No account or login is required.

The About page includes:
- Open-source attribution to VoiceSnap: https://github.com/vorojar/VoiceSnap
- License/copyright details: https://shanghai3168.github.io/privatevoice-dictation-support/
- Privacy policy: https://shanghai3168.github.io/privatevoice-dictation-support/privacy.html
- Support and feedback: https://shanghai3168.github.io/privatevoice-dictation-support/support.html

Official Apple References Checked

  • App Sandbox is required for Mac App Store distribution: https://developer.apple.com/documentation/security/app-sandbox
  • Mac App Store provisioning profile is required for manual upload signing: https://developer.apple.com/help/account/provisioning-profiles/create-an-app-store-provisioning-profile
  • Uploads use Xcode, altool, or Transporter: https://developer.apple.com/help/app-store-connect/manage-builds/upload-builds/
  • Mac App Store packaging uses signed installer packages: https://developer.apple.com/documentation/xcode/packaging-mac-software-for-distribution