Personal information
Names, email addresses, phone numbers, and identifiers.
CUSTOM POLICY SCOPEUseful AI starts with a useful prompt. Add a thoughtful privacy layer
between your sensitive information and your next LLM conversation.
Choose a fictional example and protect the prompt.
Everything in this playground stays in your browser.
Draft a follow-up for Alex Morgan at alex.morgan@example.com about the £24,500 proposal for Northstar Demo.
A useful prompt.
With sensitive details protected.
This demonstration replaces predefined fields in synthetic examples. It does not detect arbitrary PII, contact an LLM, or offer a guarantee that a prompt is safe.
Identify information governed by your policies.
Replace sensitive values with context-preserving tokens.
Use the reviewed, protected prompt in your LLM workflow.
Check the response and restore values only where authorized.
Detection, policy controls, LLM integrations, and authorized reconstruction are implementation capabilities to scope with your team. The playground illustrates masking only.
Build a protection workflow around the information
your organization handles and the way your people work.
Names, email addresses, phone numbers, and identifiers.
CUSTOM POLICY SCOPEClient references, deal values, and internal project details.
CUSTOM POLICY SCOPECredentials, access tokens, and infrastructure details.
CUSTOM POLICY SCOPEDiscuss API integration, an internal interface, or a pre-prompt step in an existing application. Deployment and data boundaries are agreed during discovery.
Define approved categories, access rules, review responsibilities, and audit metadata. Retain only the information needed for your agreed workflow.
Explore a private environment or a managed setup, subject to technical discovery. Hosting, residency, retention, and operating responsibilities need an explicit agreement.