Privacy Policy
Our Philosophy
The best way to protect your data is to never have access to it. All AI processing — speech recognition, speaker identification, LLM summarization, and semantic search — happens entirely on your device using Apple Neural Engine. We have no servers that receive, store, or process your recordings.
What We Store Locally
All data remains on your device:
- Audio recordings (M4A format with AAC encoding)
- Transcribed text with timestamps and confidence scores
- AI-generated summaries, action items, and insights
- Voice profile embeddings for speaker recognition
- Semantic search embeddings for RAG queries
- Meeting metadata (titles, duration, participants)
What We Never Collect
- Audio recordings or transcriptions — they never leave your device
- Usage analytics or behavioral data
- Device identifiers or fingerprints
- Location data
- Crash reports or diagnostics
- Any form of telemetry or tracking
Security Architecture
Six layers of protection for your data:
All files encrypted with device passcode, inaccessible when locked
Files overwritten with 0x00 bytes before removal
Face ID / Touch ID required for data export
Visual confirmation of network isolation status
iCloud Sync (Optional)
You may optionally enable iCloud sync for meeting data. When enabled, your encrypted data syncs to your personal iCloud account. We never have access to your iCloud data. You can disable this at any time in Settings.
AI Models
The app downloads AI models (~2-3GB) to your device for offline processing. These include speech recognition, voice activity detection, speaker diarization, language model, and embedding model. Models are stored separately from your data and are not deleted when you clear your meetings.
Network Policy
NSAppTransportSecurity blocks all arbitrary network loads. The app functions completely offline after initial model download. No meeting data is ever transmitted over any network.
Contact
Questions? Email privacy@resonancestudio.ai