How we work

Your data is yours. Period.

Self-Aware, Inkling, Quantum Compass, and Halo Mind Sync are built around six commitments. They aren't a footer disclaimer. They're the architecture.

The six commitments

Data sovereignty, in plain language.

1

On-device by default

All raw entries, embeddings, decision logs, outcomes, and value hierarchies are stored locally on your device. Nothing leaves the device unless you explicitly enable cross-device sync.

2

End-to-end encrypted sync

If you turn sync on, your data is encrypted on your device with a key derived from your passphrase or biometric. Our servers see opaque ciphertext, never your journal or decision content.

3

No cross-user aggregation

Your data is never used to train, calibrate, or inform any other user's model. There is no shared model trained on user content. Your weight vector and pattern history are personal and isolated.

4

No third-party analytics on content

Journal content, decision content, mood data, and free-text entries are never sent to third parties. Any usage telemetry (session length, crash reports) is anonymous, opt-in, and off by default.

5

Export and delete on demand

You can export your full data as encrypted JSON at any time. You can delete all stored data with a single action. No retention. No soft-delete with hidden recovery. When you say delete, the data is gone.

6

Biometric or passphrase lock

Self-Aware and Quantum Compass require Face ID, Touch ID, or a passphrase at every cold open. Journal-class content deserves journal-class protection. This is non-optional.

The integrity contract

What we promise about our models.

Quantum Compass and Self-Aware make claims about your life. For those claims to be trustworthy, three things must hold: the model must produce the same output from the same inputs, it must never invent facts, and it must not import biases you did not introduce.

Same question, same data, same model version. Same answer.

Every time. Deterministic by construction.

Every claim about you can be traced to data you provided.

Nothing is invented. "Show evidence" is one tap away on every output.

No demographic profiling, ever.

The app does not know your age, race, gender, religion, income, or any other protected attribute. The model literally cannot use what it does not have.

No cross-user inference, ever.

Your model is built from your data alone. There is no "people like you tend to…" reasoning anywhere in the system.

When the data is thin, the app says so instead of guessing.

"Insufficient data" is built into the contract. Fabrication is not.

You can see the math at any time.

"Show your work" exposes the five input scores, your weight vector, the formula step by step, and every record that contributed.

You can correct any data the app holds about you.

The app recomputes when you do.

You can export or delete your data at any time.

Without asking us. Without waiting on anyone.

The ethical stance

Entertainment, not clinical.

Every Divine Karma Institute application is a reflective and entertainment product. They are not medical devices, not mental health services, not diagnostic tools, and not substitutes for therapy or licensed professional care. This is not a footer disclaimer. It is the operating philosophy embedded in the language, the data architecture, the analytics, and the App Store listings of all four applications.

Clinical language is prohibited across the suite. Words like "diagnosis," "treatment," "disorder," "symptom," "patient," "cure," and "prescription" do not appear in our UI, marketing, or product documentation, except in disclaimers explaining what the apps do not do.

If you are in crisis or considering harming yourself: please contact a licensed mental health professional, call or text 988 (US Suicide & Crisis Lifeline), text HOME to 741741 (Crisis Text Line), or find local resources at findahelpline.com. Help is available, and you deserve it.
"Generative AI is never permitted to produce a value that enters the scoring formula or to make an interpretive claim about you that is not grounded in your own data." The Architectural Rule