Privacy by architecture
Your data stays yours.
Privacy-first isn’t a badge. It’s how NeuralShield is built: encrypt signal payloads on-device, minimize what leaves the phone, share only with your explicit consent — always revocable.
Encrypt locally
Feature payloads are encrypted on your device before transmission.
You control sharing
Nothing goes to family, clinicians, or anyone else without your consent.
GDPR by design
Minimization, purpose limitation, and erasure rights as product principles.
If we can’t explain it simply, we don’t collect it.
You can’t build a nervous-system baseline on fear.
Voice rhythm, typing cadence, and pulse are sensitive signals. NeuralShield is for people who want awareness without surveillance theatre — privacy is the product, not a footer link.
Clarity over legalese
| Signal | What we use | What we don’t |
|---|---|---|
| Voice | On-device acoustic features from explicit in-app check-ins (rhythm, pauses, pitch-related features). | A retained library of raw conversation recordings; always-on ambient spying on other apps’ calls. |
| Typing | Timing patterns between keypresses (cadence features only). | The words or characters you type. |
| Resting pulse | Heart-rate variability features from allowed camera check-in contexts. | Stored camera video or photo albums of your face. |
Encrypted on your phone before anything leaves
- You complete an allowed check-in or in-app capture
- Features are derived on-device
- Payload is encrypted locally
- Only then — minimal necessary transmission under your rules
Granular. Revocable. Yours.
Permissions are scoped. Care-circle sharing is opt-in, never default. You choose who sees what, and you can withdraw access without losing the idea that the product is yours first.
GDPR by design — not bolted on
Data minimization, purpose limitation, and the right to erasure are design constraints. To exercise access or deletion rights, use the contact path in our Privacy Policy (entity and process details are finalized with counsel).
Read the full Privacy Policy →What NeuralShield will not do
- Lead with anxiety-maximizing public stress scores as the product
- Sell your personal signal data for advertising
- Silently share with family or employers
- Position itself as a voice “lie detector” or interrogation tool
Different from score apps and wearables
Unlike scored vocal check-ins or ring-based recovery leaderboards, NeuralShield emphasizes a personal baseline on your phone, multi-signal context, and privacy as architecture. Early Access is where that product starts.
Privacy questions
- Is my typing content read or stored?
- No. We use timing between keypresses, not message content.
- Do you store raw audio of my conversations?
- Raw audio is not stored or transmitted. Features are derived on-device from explicit in-app voice check-ins.
- Who can see my baseline?
- You by default. Others only with explicit, granular, revocable consent.
- Can I delete my data?
- Yes — the right to erasure is a design goal. Contact us via the Privacy Policy to exercise access or deletion while self-serve tooling matures.
- Is data sold to advertisers?
- NeuralShield does not sell your personal signal data for advertising.
- Why do you need the microphone at all?
- Only for explicit voice check-ins that produce on-device acoustic features. You control permissions.
- How is this different from voice stress analysis tools?
- NeuralShield is built for your personal baseline and calm context — not one-shot stress labeling for third parties.