A calm second look at AI-written code.

Catarina checks the code ChatGPT, Claude and Gemini give you, and tells you in plain words if something looks risky, before you copy or run it.

Add to Chrome — coming soon

Free. No account. Your code never leaves your browser.

import requests
r = requests.get(url, verify=False)
print(r.status_code)

High

Certificate checks are turned off · Line 2
Someone on the same network could pose as the server.

What she looks for

Leaked secrets

API keys, tokens, private keys, passwords written into code.

Injection

SQL built from text, shell commands, unescaped user input.

Unsafe settings

Disabled certificate checks, weak hashes, open CORS, risky deserialization.

How it works

Open ChatGPT, Claude or Gemini and ask for code. Catarina outlines each block that needs a look, adds a short note with the line number, why it matters and how to fix it, and warns you if you copy a high-risk block.

Private by design

All checks run on your own computer. There is no server, no account and no tracking. Read the privacy policy.

What she doesn't do

Catarina matches known risk patterns. She doesn't find missing login checks or flaws that span several files. "Clear" means no known pattern matched, not that the code is safe.