Trust Center

Five questions.
Five answers.

Before you trust an AI with your contracts, your medical records, or your family's history, you should ask hard questions. Here are the five we insist on answering.

Your data

Where does my data actually go?

Files stay on your device by default. When you opt into cloud GPU power, only the request leaves your machine. Memory is wiped after every request. Nothing is logged. Nothing is trained on.

Audit trail

How can I prove what AskQP did?

Every meaningful action is sealed into a tamper-evident, hash-chained record. Reasoning is captured before execution. Holds up in audits, in court, and against quantum attacks that don't exist yet.

Kill switch

How do I stop everything, instantly?

Five scopes. Two modes. Under 500 milliseconds. Cannot be disabled, overridden, or bypassed. The kill switch itself is sealed into the audit trail every time you use it.

Membrane

How does AskQP reject malicious documents?

Before any document reaches your index, it passes through a checking layer that detects prompt injection, jailbreak attempts, and embedded payloads. Bad inputs get quarantined.

Disclosure

What if you find a vulnerability?

Coordinated disclosure with safe-harbor terms for good-faith research. We tell you what to expect from us, what we expect from you, and how to report a problem.

Our posture in 30 seconds

We don't ask you to trust us.

Every claim on this page can be independently verified. The architecture is documented. The crypto is open standards. The audit trail proves itself. Trust isn't a posture. It's structural.

Local-first by default

Nothing logged, nothing trained on

Post-quantum cryptography

Tamper-evident audit trail

Kill switch you cannot disable

Air-gap compatible

Read the architecture. Then decide.

The technical deep-dive is one click away. Or download AskQP and run the verification yourself.

Technical deep-dive Download AskQP

Five questions.Five answers.