AIGuard
Last updated: September 2026
At AIGuard, privacy is not just a policy—it is our core architecture. This document explains how we collect, process, and secure the data that flows through our AI proxy infrastructure.
We collect information strictly necessary to provide the service and maintain compliance audits, including:
Zero Data Brokering: We do not sell your information. Our business model relies on subscription fees, not data mining. We process data solely to execute PII tokenization and threat detection for your application.
AIGuard is designed around the principle of zero-trust architecture. Our system acts as an intermediate proxy between your application and third-party LLMs.
[NAME_001] back to "John Doe") is held securely in active memory for the duration of the request cycle and is destroyed immediately upon returning the LLM response to you.This architecture inherently prevents AIGuard from becoming a liability vector. In the event of a breach, there are no databases containing your users' raw prompts to be compromised.
While we do not store raw prompts, we do store cryptographic request metadata (such as timestamps, risk scores, and the types of PII detected) to generate PDF audit reports. This data is handled asynchronously via secure background workers.
Under GDPR Storage Limitation principles, we enforce automated data retention lifecycles based on your subscription tier:
Depending on your location, you possess rights under regional frameworks (GDPR, CCPA, HIPAA):
To exercise these rights, or to configure an Enterprise Bring-Your-Own-Database (BYOD) deployment for total data sovereignty, please contact our team at contact@aiguard.solutions.
We may update this Privacy Policy as regulatory frameworks (such as the EU AI Act) evolve. We will notify you of any material changes via your dashboard. If you have any questions regarding this policy, please reach out to support@aiguard.solutions.