PromptGuard is a security layer between an application and its LLM provider. Checks run before the request reaches the model, so a blocked prompt was never sent and a redacted value was never transmitted.
Every decision is written down.
There is no mission statement below. A reviewer opening this page wants four things — where the product sits, what evidence exists for it, where it runs and what is retained, and how a change reaches production — so those are the four things it contains, and then the limits.
Your request has been submitted successfully.
