Trust & Legal

We built your privacy in. Not written in.

Most platforms promise not to read your data. We built a system where we cannot — even if we wanted to. Here's exactly how it works.

The Big Promise
"We — the Noobs Learn AI team — cannot access, read, sell, or share your conversations, memories, or any content you share with your AI agent. Your data is yours. Period."

This is not marketing. This is how the system is built. No admin panel. No conversation viewer. No database query that surfaces your messages. The access path does not exist.

1

Privacy by Architecture

There's a difference between a company that promises not to read your data, and a company that cannot read your data. We're the second kind.

Your agent's conversations, memories, and soul contract are stored inside an encrypted, isolated instance. Our team has no admin tool, no internal dashboard, no query that can reach inside that instance. The isolation is structural — not policy-based.

What this means in practice

If a government agency issued a valid legal order demanding access to your agent conversations, we would be technically unable to comply — because the access mechanism does not exist. We would notify you of any such request to the extent permitted by law.

🔐

AES-256-GCM Encryption

All agent data stored using military-grade encryption. Accessible only to your agent instance.

🧱

Process Isolation

Each agent runs in a separate process with independent memory. No cross-instance data access is possible.

🔏

Integrity Seals

SHA-256 cryptographic seals detect unauthorized modification to your soul contract. Tampering triggers an alert.

🚫

No Admin Backdoor

No admin panel. No team viewer. No way for any team member to read your conversations. The architecture does not permit it.

2

Two Types of Data — The Distinction That Matters

We separate data into two clear categories. This is important to understand.

Category What It Is Who Can Access
Platform Data Technical metrics — response times, error logs, usage patterns. No content. Our team (aggregated)
Agent Data Your conversations, memories, soul contract, everything you share with your agent. Your agent only

Honest trade-off you should know

Because we cannot access agent data, if an agent instance is breached by an external attacker, our ability to assess what was exposed is limited by the same architecture that protects you. In any breach, we'll assume the full instance contents were potentially exposed and notify you within 72 hours — conservatively, not minimally.

What we collect for platform health — and nothing more:

WhatWhyExample
Message volumeCapacity planning"User had 47 exchanges today"
Response latencySpeed optimization"Avg response: 2.3 seconds"
Task categorySkill improvement"scheduling" — NOT the content
Error eventsBug fixing"PDF processing failed at 14:23"
Agent performance scoresQuality monitoring"Satisfaction signal: positive"
3

Your Rights — And How to Use Them

You have full control over your data. These are not just policies — they're commands you can run today.

What happens if you stop paying

Day 1–6
Agent works normally. One gentle daily reminder. Nothing restricted.
Day 7
Agent suspended. Cannot respond. Your data remains fully intact.
Day 7–37
30-day grace period. Reactivate anytime — same memory, same relationship, nothing lost. Reminders at day 15 and day 30.
Day 37+
Permanent deletion — all data is irreversibly removed. We do not want this to happen. We will remind you twice before it does.
4

Financial Advisory — Clear Limits

Some agents include financial coaching skill packs. We are transparent about what these can and cannot do.

⚠️ Not a Licensed Financial Advisor

Under Vietnamese law (Securities Law 2019, Article 90), providing investment advisory services requires professional certification. Our agents do not hold and do not claim any such certification.

Financial skill packs are strictly limited to general financial education. Our agents will never:

Prohibited ActionWhy
Predict specific prices or market movementsSpeculation — not education
Recommend specific financial productsRequires licensed certification in VN
Provide expected return percentagesMisleading without full context
Advise on securities trading strategiesOutside permitted scope

When any agent responds to financial, medical, or legal questions, an inline reminder is automatically included. This is not optional — it is enforced at the system level.

The same limits apply to medical and legal queries. Agents provide general knowledge — never professional advice. Always consult a qualified professional for critical decisions.

5

Full Legal Documents

This page is a plain-language summary. The full documents are below — written clearly, without hidden fine print.

Independent Security Audit — Pending

We are committed to having our architectural privacy and encryption claims independently verified by a third-party security auditor. Results will be made available to users upon request. Until then, these claims represent our engineering team's description of the system as built. We stand behind them and invite technical scrutiny.

6

Questions

If anything here isn't clear, contact us. We wrote these to be read — not to hide behind.

Telegram: t.me/grey_agenticAI
Email: greyai1993@gmail.com