Transparency

Trust Center

Understanding how MADDY works, how we protect your information, and the commitments we make to every person who joins us.

Why This Exists

Trust should never depend upon hidden practices.

We believe people deserve to understand how the systems they use operate. Transparency is one of the ways we practice stewardship — explaining ourselves before uncertainty becomes anxiety.

This page is a guide, not another legal document. It points you to the detailed commitments in our Privacy Policy, Terms of Service, and FAQ when you want to go deeper.

Our Public Commitments

These are the promises we make publicly. They are explained in full in our Privacy Policy and Terms.

We do not sell your personal data

We have no business model based on selling people, attention profiles, or memory materials.

We do not claim ownership of your conversations

Your information remains yours. We hold a limited stewardship permission to operate features you choose to use.

No training without explicit consent

We do not train future AI models on your conversations without your informed and explicit consent.

Privacy protects dignity

Privacy exists here because people deserve dignity — not merely because regulations require compliance.

Transparency builds trust

We explain what we collect, why, how long we keep it, how we protect it, and how you can act on it.

You remain in control

Access, correction, deletion, export, and consent withdrawal are rights we strive to honor. Ask us when tools are still catching up.

Understand these commitments

This Trust Center states what we promise. Public Education explains what those promises mean in practice — in plain language, without replacing the Privacy Policy or Terms.

Enter Public Education →

Understanding Your Information

A short map of how information is treated. Full detail lives in the Privacy Policy.

Memory & Conversations

Where companion or chat features are used, conversations exist to serve you — to provide responses and continuity you requested. They are not a silent path into model training. Live companion chat is not currently offered on the public site.

Optional memory and longer-term continuity features are part of Elysium’s vision. Where those features are not yet available on the public site, we say so plainly:

Planned for future releases.

When memory settings ship, they should give you clear control: what is stored, how long it remains, and how to delete or release what you entrusted. Consent and agency will guide those controls — not silent defaults that expand without your knowledge.

See Privacy Policy §6 and Terms §7 for the constitutional framing already in force.

AI Transparency

  • Operational processing. When companion or chat features are used, conversations are processed to generate responses and provide the experience you asked for.
  • Not the same as model training. Serving a reply is different from using your conversations to improve future AI models.
  • Optional research. If improvement or research programs are ever offered, they will require informed and explicit consent. Participation will always be optional.
  • Constitutional oversight. Development is guided by stewardship principles — dignity, consent, transparency, and human flourishing — not by engagement optimization alone.

Live companion chat is not currently offered on the public site.

Full language: Privacy Policy · Terms of Service

Security

Security exists to protect people. At a high level:

  • Encryption in transit. Public pages are intended to be served over HTTPS.
  • Authentication. Guest and future account flows are designed to keep access credentials reasonably protected; full member authentication continues to mature in Phase 0.
  • Infrastructure. We rely on established hosting and processors chosen with care for operational needs. We do not publish sensitive implementation details that would help attackers.
  • Responsible disclosure. If you discover a security issue, contact hello@maddy.life with enough detail for stewards to investigate. Please avoid public disclosure until we can assess and address the risk.
  • Future security reporting. A dedicated public security page with clearer reporting guidance is planned as the sanctuary matures.

Dedicated Security page — planned for future releases.

Our Constitutional Principles

These documents and commitments are meant to work together — the same civilization, not competing stories.

Constitutional rights and human flourishing are not marketing slogans here. They are the lenses through which Terms, Privacy, and product decisions are meant to be read.

Questions We Hear Often

Do you sell data?

No. We do not sell personal data, memories, or attention profiles. Privacy Policy §3.

Can I delete everything?

You may request deletion of personal information we hold. We honor requests as promptly as we reasonably can, subject to legal retention, security needs, and technical limits of backups. Privacy Policy §10.

Who owns my conversations?

You do. We do not claim ownership of your conversations. We hold a limited stewardship permission to operate features you use. Privacy Policy §2 · Terms §6.

Can I export my information?

Portable files you create are yours. Where server-side export exists, we will help; where Phase 0 still lacks a path, we will say so honestly. Privacy Policy §11.

Do you train on my conversations?

Not without your explicit consent. Operational processing to generate responses is different from model training. Optional improvement programs, if ever offered, require informed consent and remain optional. Privacy Policy §6.

What happens if Elysium changes?

Meaningful privacy changes deserve clear communication — not merely revised legal text. Transparency should grow alongside our technology. Privacy Policy §13.

Accountability

If we fail to live up to these commitments, we want to know. Stewardship includes accountability.

Write to hello@maddy.life or use our Contact page. A human steward will read your message.

Closing

Technology changes quickly. Principles should change carefully.

Our goal is not merely to build useful software. It is to build software worthy of long-term trust.