Privacy

llm11 is a proxy: your request passes through our infrastructure to route and verify it, which means we see it in flight. By default we store only metadata about each request - which model answered, token counts, cost, latency, which checks ran and their verdicts - never the prompt or response text itself. Storing content is an explicit, per-project opt-in with a retention window you choose, and it exists only to support features that need it, such as reviewing what a groundedness check actually flagged.

Content is never used to improve routing or verification for any other customer. Model providers we route to (via Vercel AI Gateway or OpenRouter) and, where configured, TypeSafe AI for triage, see the request under their own respective data-handling terms.

This is a working draft, published so the posture is stated rather than assumed. It will be replaced with counsel-reviewed terms before this product accepts production traffic from customers outside its design-partner group.