Our privacy-first approach
CalmSignal processes everything 100% on your device—your text never leaves your device.
Overview
CalmSignal is designed with privacy at its core. All text sentiment analysis runs 100% on your device using on-device AI models. We never upload raw messages or keyboard input to servers. Everything is processed locally for complete privacy.
How on-device processing works
CalmSignal uses different on-device AI technologies depending on your platform:
All platforms use a privacy-first inference cascade: the system tries the most accurate on-device model first, falls back to heuristics if needed, and only uses backend APIs if you explicitly opt-in (disabled by default).
- iOS: CoreML models that run on the device's Neural Engine
- Android: ONNX models or heuristic analysis that runs locally
- Browser Extensions: ONNX Runtime Web (WASM) or heuristics—all in your browser
What this means for you
Your text is analyzed entirely on your device. No raw text, messages, or keyboard input is ever sent to our servers. Optional cloud sync (disabled by default) only syncs aggregated insights if you explicitly enable it, and even then, your raw text never leaves your device.