Runs on your device

Use AI freely.
Share nothing you shouldn't.

GuardAI reads what you type into AI tools and catches sensitive details before they leave your browser. Everything runs locally, so we physically cannot see your data.

Free forever No account Zero external calls
chatgpt.com
YOU

Hi, my name is Sarah Chen and my number is 0412 345 678 if you need it. Can you draft a reply?

Sensitive data caught Before send
Name Sarah ChenEmma Walsh
Phone 0412 345 6780498 111 222

Works on the AI tools you already use

New platforms are added as they launch

01 / Local

Zero external calls. Nothing is sent anywhere to be processed.

02 / Anonymous

No account required. There is nothing to sign up for.

03 / Open

Nothing hidden. The extension is open to inspection.

Demo · 45 seconds

Watch it work on a real message

One message goes in. GuardAI swaps the name, the number and the address, sends the safe version to ChatGPT, then puts your real details back into the answer.

Full demo · 0:45 · no sound needed

detection → masking → restore

The problem

You're sharing more than you think

90%

of people don't trust AI with their data, and use it anyway

43%

have already left ChatGPT over privacy concerns

34.8%

of prompts contain sensitive information

Every message you send to ChatGPT, Claude or Gemini can be stored, used for training and kept indefinitely. Most people find out what they shared long after they sent it.

How it works

Protection in three steps

STEP 01

You type normally

Write your message in ChatGPT, Claude or any other AI tool. Nothing about the way you work has to change.

STEP 02

GuardAI rewrites it before it sends

Sensitive details are swapped for realistic stand-ins. A different name. A different number. A different address. The AI only ever receives the fake version.

STEP 03

You get the real answer back

When the reply comes back using the fake details, GuardAI restores your real ones first. You read the real answer. The AI never had it.

Prefer to stay in control? Three other ways to run it.

Warn onlyGet alerted and decide for yourself.
Edit masksChoose exactly what gets replaced.
Send anywaySkip protection for one message.
Features

Everything you need. Nothing you don't.

Smart detection

Names, phone numbers, emails, addresses, Medicare numbers, TFNs, passwords and more.

Data masking

Real values are replaced with convincing stand-ins before the message leaves your browser.

Auto-restore

Replies come back rewritten with your real details, so the protection stays invisible to you.

Manual control

Highlight anything the detector missed and mask it yourself. You decide what gets protected.

Why you can believe it

We physically cannot see your data.

GuardAI runs inside your browser. Your messages never leave your device. There are no servers processing your text, no account holding your history and no company reading your conversations. Not won't. Can't.

End to end local Processing happens in the page you are already on.
No servers There is no backend for your text to travel to.
Zero data collection No database, no analytics, nothing kept about you.
FAQ

Common questions

Yes, completely free. No account, no subscription and no hidden costs.

GuardAI uses pattern matching that runs entirely in your browser. Nothing is sent to a server to process your messages.

ChatGPT, Claude, Gemini, Copilot, Perplexity and more. We keep adding support for new platforms.

None. There is no server, no database and no analytics on your messages. The only thing stored is your local masking table, which lives on your device and can be cleared whenever you want.

Start protecting your AI conversations.

Free, instant and no account needed.

Add to Chrome

Used by privacy-conscious professionals across Australia