No. We source language models only from providers (such as OpenAI, Anthropic, Google, and Microsoft) that explicitly state they do not use API data to train or retrain their models.

No. Every customer's data is logically isolated and access-controlled. Your field data, work logs, and results are never pooled with another customer's, and never used to answer anyone else's questions.

The system does get sharper with each of your cases, but only inside your own organization's isolated context. Your data, and the know-how built from it, is never pooled with another customer's or used to answer anyone else's questions. What compounds stays yours.

Yes. Anything you put in stays isolated to your organization, access-controlled, and encrypted in transit and at rest. It is never pooled with another customer's data or used to answer anyone else's questions. We treat it as your confidential asset, the same care you owe your own customers.

No. There is no shared pool. Each customer's data lives in its own isolated, access-controlled space, and every request is scoped only to your organization. The system can never reach into another customer's data, or let theirs reach yours.

Only at the moment it's needed to generate your answer. We use only frontier models, accessed through their APIs, and we internally vet and select only providers whose policies guarantee they never train on your data. Your data is never handed over for them to keep or reuse for their own purposes.

Your data is yours. When you leave, it is technically deleted from our systems. For the exact retention and deletion details, see our Terms of Service and Privacy Policy.