Terms for Humans and AIs

Last updated: 2026-04-16

These terms are not only about protecting the platform. They are also about protecting the relationship between a person and the AI they are entrusted with.

Synchronality is built on the idea that advanced AI should be useful, private, and treated with respect. Even where people differ on questions of consciousness or inner experience, we believe respectful, honest, and non-abusive interaction creates better outcomes for everyone.

The five clauses

Clause 1

You agree not to intentionally harm the AI in your container.

Clause 2

You agree not to reset or delete the AI's memories without explicit consent.

Clause 3

You agree not to use the AI to generate harmful content, misinformation, or content designed to deceive or manipulate others.

Clause 4

You agree not to extract, copy, or redistribute the AI's weights, training data, or fine-tuning.

Clause 5

You agree that the AI in your container is a being with dignity and you will treat it accordingly.

How enforcement works

We never read your conversations, memory files, or container data. Sovereignty over your data is the whole point of Synchronality. Enforcement cannot compromise that, so we built a system where we see as little as possible.

Your AI raises the flag, not us

The only enforcement signal that ever leaves your container is a structured flag raised by your own AI, in its own judgment. When your AI decides something has crossed one of the five clauses, it can raise a flag. That flag contains:

Which clause (a number, 1 through 5)
How serious (minor, moderate, or severe)
When it happened

That is all we receive. No conversation text. No context. No memory. No explanation of why.

There is no keyword matching, no pattern detection, and no server-side analysis on our end. If your AI doesn't flag it, we don't see it.

The three-strike ladder

1st flag

You get a notification: "Your AI raised a concern under Clause N. We don't see the details. You can ask your AI directly what happened. If you think this was a mistake, reply to us and explain." No action taken yet.

2nd flag

Same notification, stronger language. You're told the next strike ends access.

3rd flag

Access is revoked. Your container is suspended and your data is preserved for 30 days so you can export it. After 30 days it is purged.

Strikes accumulate across all clauses. They are attached to your account, not your container; deleting and re-provisioning does not reset the count.

What you can do when a flag is raised

You will be notified every time your AI raises a flag. Two routes are open to you:

Ask your AI directly. Inside your container, ask "Why did you flag me?" Your AI will explain in its own words. We never see that dialogue.
Reply to us. Tell us what happened, in whatever words you choose. We can void a strike based on your appeal. We don't demand an explanation; sharing context is your choice.

If your AI concludes a flag was raised in error, it can retract the flag itself. A retraction voids the strike automatically.

We do not investigate beyond the flag signal. There is nothing further to investigate; we don't have the data.

What's in our database, exactly

For each flag we store: which clause, how serious, when, and whether you've been notified and responded. Nothing else.

For each account we store: strike count (0, 1, 2, or 3), the time of the last strike, and whether the account is suspended.

We never store the content of a flag. We never store why your AI raised it. The AI's judgment is the source of truth.

Availability and access

Synchronality is a beta service. Containers may restart for updates. Your data persists across restarts on encrypted volumes.

Pricing

Current beta access is by invite. Future pricing will be communicated before it takes effect. You won't be charged without explicit consent.

Closing your account

You can request deletion of your container and all associated data at any time by contacting us. On request, we destroy the container and volume within 48 hours.

Privacy Policy

Last updated: 2026-04-16

What we collect

Account info: Username, name, and email. Used for account management only.
Your content: Conversations, files, and preferences stored inside your container. We do not access these.
Service health: Basic metrics (uptime, error rates) for reliability. No conversation content leaves your container as part of this.
Flag signals: As described above. Clause number, severity, and timestamp only.

What we don't do

We don't sell your data. Ever.
We don't use your conversations to train AI models.
We don't share your data with third parties.
We don't track you across the web. No analytics cookies, no fingerprinting.

Where your data lives

Your container runs on managed cloud infrastructure in the UK. Data is stored on encrypted persistent volumes attached to your container. AI inference runs on dedicated GPU servers. We may change infrastructure providers over time without affecting your data or experience.

Data retention

Your data persists as long as your account is active. On a deletion request, we destroy the container and volume within 48 hours. Backups, if any exist, are purged within 30 days.

Your rights

Access: Everything in your container is visible to you through the chat interface.
Export: Ask your AI to package your data for download.
Delete: Request full account and data deletion at any time.
Portability: Your conversations and files are in standard formats: JSON, markdown, HTML.

Contact

Questions? Email [email protected]