Security

Effective date: 27 September 2026

Applies to unlimitless.ai and the Unlimitless service at api.unlimitless.ai

Your reasoning is the most valuable thing you will put into Unl, so protecting it is the core of the product, not a feature around it. These are our commitments, and every one of them holds today.

Your data stays yours

Unl keeps only what you choose to save. That means the decisions you make, the reasoning behind them, the targets you commit to and the thoughts you keep.

When your AI reads one of your own tools through Unl, the result goes to your AI in the moment it asked. Unl holds your criteria, not your tools’ data.

We never read your content to police it. Nothing you save is scanned, moderated or mined, and we never train models on it. Nothing you save reaches analytics or error tracking either. Those systems can see that a save happened, never what it said.

Isolated by design

Every workspace is sealed. Every read and every write is checked against your workspace twice, by two independent protections, so what you save is reachable only from inside it.

What you save is kept apart from account and usage records, and the two never touch.

No agent can write without you

Any AI you work with can read what bears on its work. Only you can change what is kept. No model, message, web page or tool result can write to your world, however it is worded and whatever instructions it carries.

Unl only ever takes in what is reported to it. It never reaches into a model or an agent to inspect or control it.

Nothing is overwritten. When you change your mind, the new position is kept beside the old one, so you can always go back to what you thought before and why.

Your tool keys stay sealed

Keys you connect for your own tools are encrypted at rest with AES-256. They are never logged, never returned once saved and never shown to any model. Removing a key destroys it.

You can leave with everything

Export your whole workspace as machine-readable JSON from your portal at any time, and ask for its deletion from the same place. There is no lock-in.

Proven on every change

Every change to Unl is blocked from release unless these protections are proven to still hold. An independent security review is planned, and this page will carry its date once it is complete.

Reporting a vulnerability

If you find a vulnerability, write to hello@unlimitless.ai with enough detail to reproduce it. A confirmed report gets a response within the hour.

We will never threaten you for telling us. Give us a reasonable window to fix it before you publish, and we will tell you when it is closed.