# Mirage

> Mirage protects individuals and organizations who want to use AI without exposing sensitive information. It acts as a privacy layer for LLM interactions, performing reversible redaction of sensitive data (PII) — names, emails, phone numbers, financial figures, organization names, and secrets — before it reaches cloud AI providers. Tokens in the response are automatically rehydrated so the user sees original values. Your data never leaves your machine unmasked.

Mirage is available as a Chrome extension and a desktop app. The extension supports Claude and ChatGPT, with LeChat coming soon.

- All detection and redaction happens locally on your device — no external API calls
- Automatic rehydration replaces tokens in LLM responses back to original values
- Free

## Pages

- [Home](https://getmirage.co): Main landing page with animated hero showcasing the redaction concept
- [Why Mirage](https://getmirage.co/why-mirage): Educational article comparing Mirage to VPNs and explaining why a privacy layer for AI is essential
- [Privacy Policy](https://getmirage.co/privacy): Privacy policy — Mirage never collects, stores, or transmits sensitive data
- [Terms of Service](https://getmirage.co/terms): Terms of service for using Mirage

## Optional

- [Privacy Policy](https://getmirage.co/privacy): Full legal privacy policy details
- [Terms of Service](https://getmirage.co/terms): Full legal terms of service details