📱 iOS App — On-Device AI Disclosure
VoicePad AI for iOS uses the open-source Whisper model (originally released by OpenAI) running entirely on your device via whisper.cpp. This is a locally-executed model — no audio, voice data, or transcribed text is ever transmitted to OpenAI, any cloud service, or any third party.
VoicePad AI does not use the OpenAI API, ChatGPT, or any internet-connected AI service. All speech recognition happens in your device's local memory and is discarded immediately after processing.
Data sent to third parties: None. The app functions fully offline after the one-time license activation.
🤖 Android App — On-Device Processing & Accessibility Services
VoicePad AI for Android runs the open-source Whisper speech recognition model entirely on your device via whisper.cpp. No audio, voice data, or transcribed text is ever transmitted to any cloud service or third party.
Android Permissions — Full Disclosure
VoicePad AI declares the following Android permissions. Each is used only for the purpose stated below:
| Permission | Purpose | Data Handling |
|---|---|---|
| RECORD_AUDIO | Capture your voice while you actively dictate | Processed in volatile RAM only; discarded immediately after speech recognition; never stored, never transmitted |
| BIND_ACCESSIBILITY_SERVICE | Power the optional VoicePad Direct feature: type your dictated text into the active text field of another app | See "Accessibility Service Disclosure" below |
| SYSTEM_ALERT_WINDOW | Display the optional Float Orb — a draggable button that lets you start dictation from any screen | Renders only the Orb itself; reads no content from other apps |
| FOREGROUND_SERVICE + FOREGROUND_SERVICE_MICROPHONE + FOREGROUND_SERVICE_SPECIAL_USE | Keep the recording session and Float Orb alive while you dictate, as required by Android | Required by Android OS; no data is collected |
| POST_NOTIFICATIONS | Display the foreground service notification while recording (Android 13+) | No data is collected |
| INTERNET + ACCESS_NETWORK_STATE + ACCESS_WIFI_STATE | One-time license activation, optional update checks, and WiFi sync (LAN-only, opt-in) | Used only for license validation and your own LAN sync; no audio or transcribed text is ever sent over the internet |
| DOWNLOAD_WITHOUT_NOTIFICATION | Silent download of Whisper speech recognition model files on first run | Models are downloaded directly from voicepad.tech servers; no user data is sent |
| WAKE_LOCK | Keep the device awake during dictation | No data is collected |
| VIBRATE | Provide haptic feedback when dictation starts/stops | No data is collected |
| BILLING | Google Play Billing for optional Pro upgrade | Standard Play Billing flow; no usage data shared |
| RECEIVE_BOOT_COMPLETED | Restore Float Orb after device reboot (only if you previously enabled it) | No data is collected |
| READ_MEDIA_AUDIO | Access exported dictation audio files in your local Music folder when you choose to play them back | Local file access only |
Legacy storage permissions (READ_EXTERNAL_STORAGE, WRITE_EXTERNAL_STORAGE) are declared with maxSdkVersion caps for backward compatibility with older Android versions only and are not active on modern devices.
Accessibility Service Disclosure (VoicePad Direct)
The VoicePad Direct feature uses Android's BIND_ACCESSIBILITY_SERVICE permission. This is a sensitive permission, and we treat it accordingly. The following is a complete and accurate description of how VoicePad's accessibility service operates:
What the accessibility service DOES:
- Types your dictated text into the currently focused text field of the app you are actively using
- Activates only when you press the VoicePad recording button
- Operates only while you are speaking; injects nothing during silence
What the accessibility service DOES NOT DO:
- Does NOT read screen content
- Does NOT capture, log, or analyze keystrokes from any app
- Does NOT monitor user behavior, browsing, messages, or app usage
- Does NOT collect, store, log, transmit, or share any data observed through the AccessibilityServices API
- Does NOT contain any code that sends data over the internet
- Does NOT run when VoicePad Direct is disabled
User control:
- VoicePad Direct is OFF by default on every fresh install
- You must explicitly enable it in Android system settings (Settings → Accessibility → VoicePad AI)
- Before requesting accessibility permission, VoicePad shows a prominent in-app disclosure screen explaining what the service does and links to this Privacy Policy
- You can revoke accessibility permission at any time in Android system settings
Data sent to third parties from accessibility usage: None. The accessibility service contains no network code.
Float Orb Disclosure
The Float Orb feature uses SYSTEM_ALERT_WINDOW to display a small draggable recording button on top of other apps. The Orb itself reads no content from other apps. It is OFF by default and can be disabled at any time in VoicePad settings or in Android system settings (Settings → Apps → VoicePad AI → Display over other apps).
WiFi Sync (Android)
VoicePad's optional WiFi Sync feature allows you to sync your dictation history between your VoicePad devices on the same local network. It uses mDNS (_voicepad._tcp) for discovery and HTTP on port 7777 for transfer. This feature is opt-in, off by default, and operates entirely on your local network — no data is sent over the internet.
🖥️ macOS App — On-Device Processing & System Permissions
VoicePad AI for macOS runs the open-source Whisper speech recognition model entirely on your Mac via whisper.cpp. No audio, voice data, or transcribed text is ever transmitted to any cloud service or third party.
The macOS version is distributed directly from voicepad.tech as a notarized Developer ID signed application (Apple Team ID: D8X987CDUV). It is not distributed through the Mac App Store. Direct distribution allows the global hotkey and text-injection features that the App Store sandbox does not permit.
macOS Permissions — Full Disclosure
VoicePad AI for macOS requests the following system permissions on first use. Each is used only for the purpose stated below:
| Permission | Purpose | Data Handling |
|---|---|---|
| Microphone | Capture your voice while you actively dictate | Processed in volatile RAM only; discarded immediately after speech recognition; never stored, never transmitted |
| Accessibility | Power text-injection features: type your dictated text directly into the active text field of any macOS app | See "Accessibility Permission Disclosure" below |
| Input Monitoring (when applicable) | Listen for the global hotkey that activates dictation from any app | Detects only the configured hotkey combination; does not log keystrokes |
| Local Network (optional, opt-in) | WiFi Sync feature — transfer dictation history between your VoicePad devices on the same local network | LAN-only; no data leaves your local network; opt-in feature, off by default |
Accessibility Permission Disclosure (macOS)
VoicePad AI for macOS uses the CGEventTap API to type your dictated text into the focused text field of the app you are currently using. macOS treats this as an Accessibility permission request because event injection is a sensitive capability.
What VoicePad's accessibility usage DOES:
- Synthesizes keyboard events to type your dictated text into the currently focused text field
- Activates only while you are speaking; injects nothing during silence
- Is OFF by default; you must grant Accessibility permission in System Settings → Privacy & Security → Accessibility the first time you use the feature
What VoicePad's accessibility usage DOES NOT DO:
- Does NOT read screen content
- Does NOT capture, log, or analyze keystrokes from any app
- Does NOT monitor user behavior, browsing, messages, or app usage
- Does NOT collect, store, log, transmit, or share any data observed through system APIs
- Does NOT contain any code that sends data over the internet from event tap activity
User control:
- VoicePad shows a clear in-app explanation before requesting Accessibility permission
- You can revoke Accessibility permission at any time in System Settings → Privacy & Security → Accessibility → uncheck VoicePad AI
Software Updates and Notarization
The macOS app is signed with Apple Developer ID and notarized by Apple before distribution. Updates are delivered as new signed/notarized DMG installers from voicepad.tech. The app contains no auto-update telemetry that transmits data to us about your usage.
Payment and Licensing on macOS
The macOS version is licensed via LemonSqueezy (one-time Pro purchase). Email and license key are processed by LemonSqueezy and stored in our Supabase license database. See Section 5 — Third-Party Services for details. No usage data, audio data, or transcribed text is ever sent to LemonSqueezy or Supabase.
🪟 Windows App — On-Device Processing & Permissions
VoicePad AI for Windows is a Python + PyQt6 desktop application that runs the open-source faster-whisper speech recognition engine entirely on your computer. No audio, voice data, or transcribed text is ever transmitted to any cloud service or third party.
The Windows version is distributed directly from voicepad.tech as a code-signed installer (Inno Setup, signed with Certum EV Code Signing certificate).
Windows Permissions and System Access
Windows does not use a permission-prompt model like Android or iOS. VoicePad AI for Windows accesses the following system resources for the purposes described:
| System Access | Purpose | Data Handling |
|---|---|---|
| Microphone | Capture your voice while you actively dictate | Processed in volatile RAM only via PyAudio; discarded immediately after speech recognition; never stored, never transmitted |
| Keyboard input synthesis (SendInput API) | Type your dictated text directly into the focused text field of any Windows app | Synthesizes keystrokes only with your dictated text; does not read screen content, does not log keystrokes from other apps, does not contain network code |
| Local Network (optional, opt-in) | WiFi Sync feature — transfer dictation history between your VoicePad devices on the same local network | LAN-only via Zeroconf mDNS (_voicepad._tcp) and a local HTTP server on port 7777; no data leaves your local network |
| Local Disk (%APPDATA%\VoicePad) | Store local dictation history, settings, and Whisper model files | Stored in your user profile only; never transmitted |
Crash Reports
If VoicePad encounters an unexpected error, a crash report is written to %APPDATA%\VoicePad\crashes on your local computer. Crash reports are stored locally only and are never transmitted to us or to any third party. You may review or delete them at any time.
Telemetry and Analytics
VoicePad AI for Windows contains no telemetry, no analytics, and no usage tracking. We have no visibility into how you use the app.
Updates
Updates are delivered as new code-signed installers from voicepad.tech. The application does not phone home to check for updates by default.
1. Voice Data — Your Audio Never Leaves Your Device
VoicePad AI uses OpenAI Whisper, an open-source AI model that runs 100% locally on your device:
- Audio Capture: Your microphone captures audio only when you activate dictation
- Local Processing: Whisper AI processes audio entirely in your device's RAM
- Text Output: Converted text goes to your clipboard or active application
- Immediate Deletion: Audio data is discarded from memory immediately after processing
- No Network Transmission: Audio data never touches the internet — the app works fully offline
2. What We Collect — Only Your Email for Licensing
We collect the absolute minimum data required to provide you with a license:
| Data | Purpose | Legal Basis (GDPR) | Retention |
|---|---|---|---|
| Email Address | License delivery and management via LemonSqueezy | Art. 6(1)(b) — Contract fulfillment | Duration of license ownership |
| License Key | One-time validation at activation via Supabase | Art. 6(1)(b) — Contract fulfillment | Duration of license ownership |
| Transaction ID | Payment verification, refunds | Art. 6(1)(c) — Legal obligation | 7 years (tax compliance) |
After initial license activation, VoicePad AI operates fully offline. No ongoing data transmission occurs.
3. Data We Never Collect
- ❌ Voice recordings or raw audio data
- ❌ Transcribed text content
- ❌ Keystrokes, screen content, or clipboard data
- ❌ Documents, files, or user-generated content
- ❌ Location or GPS data
- ❌ Browsing history or internet activity
- ❌ Contact lists, calendar, or personal files
- ❌ Analytics or usage tracking data
- ❌ Device identifiers or fingerprints
4. Data Controller Information
Data Controller: VoicePad AI
Filderstadt, Germany
Contact: contact@voicepad.tech
Website: https://voicepad.tech
5. Third-Party Services
| Service | Purpose | Data Shared | Location |
|---|---|---|---|
| LemonSqueezy | Payment processing, license delivery | Email, transaction data | EU/US |
| Supabase | License validation (one-time at activation) | License key | EU |
| OpenAI Whisper | Speech recognition | None — runs locally on your device | On-device only |
| WiFi Sync (mDNS, port 7777) | Optional local network sync between your VoicePad devices | None — operates entirely on your local network | On your devices only |
No Analytics: VoicePad AI does not use Google Analytics, Plausible, or any other tracking service. We have no visibility into how you use the app.
6. Data Security
We implement appropriate security measures:
- HTTPS/TLS encryption for all network communications
- License data encrypted locally on your device
- Minimal data collection reduces attack surface
- No cloud storage of user data
7. Your Rights Under GDPR
As a data subject under GDPR (EU) 2016/679, you have the following rights:
| Right | Description |
|---|---|
| Access (Art. 15) | Request confirmation of processing and access to your data |
| Rectification (Art. 16) | Correct inaccurate or incomplete data |
| Erasure (Art. 17) | Request deletion ("right to be forgotten") |
| Restriction (Art. 18) | Limit processing in specific circumstances |
| Portability (Art. 20) | Receive data in structured, machine-readable format |
| Objection (Art. 21) | Object to processing based on legitimate interests |
| Lodge Complaint (Art. 77) | File complaint with supervisory authority |
To exercise your rights, email: contact@voicepad.tech. We respond within 30 days.
8. Children's Privacy
VoicePad AI is not intended for individuals under 16 years of age. We do not knowingly collect personal data from children under 16.
9. Changes to This Policy
We may update this Privacy Policy to reflect changes in our practices. Material changes will be communicated via the "Last Updated" date on this page. Continued use after changes constitutes acceptance.
10. Supervisory Authority
You have the right to lodge a complaint with a data protection supervisory authority: