DataGuard sits between your application and any LLM API — intercepting sensitive data, enforcing your compliance policy, and generating a full audit trail in real time. Before anything reaches the model.
Two-stage detection pipeline identifies sensitive entities in every prompt and every response — emails, phone numbers, payment data, legal references, medical terms, and more.
Per-client policy rules determine whether each entity is allowed through, redacted before transmission, or blocked entirely. Configurable without writing code.
Every enforcement decision is logged with a human-readable explanation — what was detected, what rule applied, what action was taken. Compliance evidence, not just logs.
Your app sends a prompt. DataGuard intercepts it before it reaches the LLM.
The classifier detects sensitive entities in the input — PII, financial data, legal terms.
The policy engine applies your rules — allow, redact, or block each entity type.
Output is scanned again. Every decision is logged with full explainability.
KERA deployed a live AI receptionist for The Lime Tree Hotel — handling real guest enquiries on WhatsApp in real time. Every guest interaction passes through DataGuard: booking references, email addresses, and phone numbers are classified and governed before reaching the language model. The deployment is live, in production, and handling real guest communications.
Simple, transparent pricing. No "contact us for a quote." No surprises.
All plans include a one-time setup fee. DataGuard can be live in under a week.
Tell us about your use case. We'll be in touch within one business day.
We'll be in touch within one business day.