Privacy Policy
Last updated: March 16, 2026
Whisperer is designed from the ground up to keep your data private. All speech processing happens entirely on your device. We do not collect, store, or transmit any personal data.
Zero personal data collected
Audio never leaves your device
No analytics or telemetry
No sign-up or login required
What Data Whisperer Processes
Whisperer processes the following data, all of which remains on your device:
- Microphone audio — Captured in real time at 16 kHz for speech-to-text conversion. Audio is processed in memory and is never written to disk or transmitted over the network.
- Transcription text — The output of the speech recognition model. On macOS, text is injected at the cursor position or saved locally for meeting transcripts. On iOS, text is stored in a local SwiftData database on your device.
- App preferences — Settings such as language selection, model choice, and shortcut configuration are stored locally via UserDefaults.
On-Device Processing
All speech recognition is performed locally using whisper.cpp, an open-source C/C++ implementation of OpenAI's Whisper model. The model runs on your device's CPU and GPU (via Metal acceleration). No audio data is sent to any server for processing.
Optional AI Refinement
MacWhisperer includes an optional AI transcript refinement feature. When enabled by the user, meeting transcripts may be sent to a third-party AI provider (OpenAI, Anthropic, or a local Claude Code CLI instance) for post-processing. This feature is:
- Disabled by default
- Requires explicit user opt-in
- Only sends transcript text (not audio)
- Can be configured to use a fully local provider (Claude Code CLI)
When using a cloud-based AI provider, the transcript text is subject to that provider's privacy policy.
Model Downloads
When you first select a Whisper model, the app downloads the model file from Hugging Face. This is the only network request the app makes for core functionality. After download, the model is cached locally and no further network access is required.
Permissions
Whisperer requests the following system permissions:
- Microphone — Required to capture audio for transcription.
- Accessibility (macOS only) — Required to inject transcribed text at the cursor position in any application.
These permissions are used solely for the stated purposes and can be revoked at any time in System Settings.
Third-Party Services
Whisperer does not integrate any third-party analytics, advertising, or tracking services. There are no SDKs that collect user data. The app does not contain ads.
Children's Privacy
Whisperer does not knowingly collect any personal information from anyone, including children under 13.
Open Source
Whisperer is fully open source. You can audit the entire codebase to verify these privacy claims at github.com/stonedMoose/iWhisperer.
Changes to This Policy
If this privacy policy is updated, the changes will be reflected on this page with an updated "last updated" date. Since the app does not collect contact information, we cannot notify users directly.
Contact
If you have questions about this privacy policy, you can open an issue on the GitHub repository.