An API that tells an AI agent what its owner has actually decided — and refuses to invent the rest.

Your data

You are trusting us with your memory. Here is what we do — and never do — with it.

• —We never train AI or machine-learning models on your data.
• —We never sell it to third parties.
• —No advertising, no profiles for targeted marketing.
• —We never access or analyze your source code.
• —Tenant isolation is enforced at the database level, with row-level security.
• —Delete your account and all data goes with it — permanently, at any time.

One key, one mind. A key is the boundary: every action is scoped to the records written under that key, and an answer can only ever cite this tenant’s decision records. Nothing crosses between tenants, because there is no path that would let it.

The honest part, said plainly: content you submit is processed by the AI providers we route to, under their no-training API terms — that is how the work gets done. It is never used to train anyone's models and never sold. And by design, much of the work is done by deterministic code that no AI model ever sees — your content reaches an AI provider only when judgment is actually needed. The full commitments live in the privacy policy — this box is the plain-language version of that page, never a substitute for it.

On the roadmap: an option to keep your content on your own machine entirely — data minimization is how this system was designed from day one, not a feature added later.