I use AI chats a lot when coding and analyzing logs.
A few days ago I almost pasted a real API key into ChatGPT while sharing some logs.
I noticed it just before sending.
But it made me realize something:
It's extremely easy to accidentally leak sensitive data when using AI chats.
So I built a small Chrome extension called PasteSafe.
What it does
When you paste text into AI chats, it scans the content and detects things like:
- API keys
- emails
- phone numbers
- IBAN
- UUID
- URLs
If something sensitive is detected, it automatically masks the values before sending.
Example:
API_KEY → [API_KEY#1]
Works with
- ChatGPT
- Claude
- Gemini
Privacy first
Everything runs locally in the browser.
- no servers
- no tracking
- no data collection
Try it
https://pastsafe-ext.github.io/pastesafe/
GitHub repo:
https://github.com/pastsafe-ext/pastesafe
Curious:
Have you ever accidentally pasted something sensitive into an AI chat?
Top comments (1)
One thing that surprised me while testing this:
emails and IDs appear in pasted logs much more often than API keys.
Curious what others accidentally paste into AI chats.