Check what you send to AI, before you send it
EraseAI Firewall checks every chat in the AI apps you choose on Android, and your messages and attached files on Chrome. If it finds personal data, card numbers, passwords or keys, it holds the message and shows you what it found.
Start free. Both open their store listing: Google Play, and the Chrome Web Store.
What it does on each device
On Android
- Protects the AI apps you choose, such as ChatGPT, Claude and Gemini.
- Checks what you type before it is sent, and holds the message if it finds something.
- You choose: Cancel, Sanitize & Send, or Send Anyway.
- An optional EraseAI keyboard, and EraseAI Safe to clean files before you attach them.
On Chrome
- Free: every message to ChatGPT, Claude and Gemini is checked on your device.
- Personal: reads inside attached PDF, Word, Excel, slides and images, and can download a clean copy.
- Personal: one-click Sanitize & Send for the message itself.
- Chrome only today. Other browsers are coming soon.
EraseAI Personal
$5 a month
Or $54 a year. Attachment scanning and one-click Sanitize & Send on Chrome, and full protection for your AI chats on Android.
Try it first, for free.
The Chrome extension is free to install, and the Android app has a free trial of 7 days or 25 messages. Subscribe when you are ready.
Before you install, here is what you should know
Plain answers about what it reads, where your text goes, and who is behind it.
What it can see
On Android, EraseAI uses the Accessibility service for one purpose: to check text you type in the AI apps you chose. Password fields are skipped, and it is not a VPN. On Chrome, it runs only on ChatGPT, Claude, Gemini and Replit.
Where your text goes
Without an account, checks run on your device and nothing is sent. With an account, message text is checked by our API over HTTPS. On Chrome, images are read inside your browser and never uploaded. See the privacy policy.
What it can't do
Detection is rules-based: patterns plus checks such as the card-number checksum. It can miss a secret it has no rule for, and it can flag harmless text. You always see the result before anything is sent.
Who makes it
Vantward Solutions Pte. Ltd., a company registered in Singapore (Reg. No. 202606980C), 68 Circular Road #02-01, Singapore 049422. The Chrome Web Store listing shows the same company.
Why we built it
I spent 17 years in security: in the Air Force, across physical, national and cyber security, and then as an aerospace engineer, where protecting sensitive information was my core duty. Then I started a software company and wrote the code by hand.
When we moved to AI-driven development, I saw people on teams paste company information into public AI tools, something that should never have happened. Then one morning, as a father, I realised I had done the same with information about my own child.
That is when my understanding changed, and EraseAI Firewall grew out of it. It is live, not a pre-product, and bootstrapped.
Founder, Vantward Solutions Pte. Ltd., Singapore