PrivacyAI Shield by Rudra

PrivacyAI Shield is a local-first browser privacy layer that helps protect sensitive information before it reaches AI chat services.

The add-on detects and sanitizes personally identifiable information (PII), credentials, and other sensitive text from prompts before they are submitted to supported AI chat interfaces such as ChatGPT, Claude, Gemini, Perplexity, Copilot, Poe, and local AI tools.

Key features:

Automatically intercepts AI chat prompts before submission.
Redacts sensitive data such as email addresses, phone numbers, API keys, access tokens, SSNs, credit card-like numbers, medical IDs, names, organizations, and locations using local AI processing on your device. The transformation happens entirely on your PC, preserving the original intent of your prompt so the experience feels unchanged. Note: response restoration is not yet fully implemented.
Uses local-first detection and sanitization so sensitive text can be processed in the browser before being sent to an AI model.
Supports local AI provider workflows such as Ollama and LM Studio.
Preserves useful prompt structure while replacing sensitive values with safe dummy data.
Works across popular AI chat websites and local development/demo environments.
Designed for privacy-conscious users, developers, security teams, and anyone who wants an extra protection layer when using AI tools.

Learn more and give report any bugs and addition wanted to add at:
GitHub: https://github.com/ru-aish/PrivacyAI
NPM: https://www.npmjs.com/package/privacyai

PrivacyAI Shield is especially useful when working with support messages, logs, code snippets, credentials, customer details, internal notes, or any prompt that may accidentally contain private information.

The extension is local-first and built to reduce accidental data exposure when using AI chat products.

Just plug it and forgot it you will never fill any changes.