Hey Product Hunt 👋 I'm Dario.

I built API Key Leak Intercept after nearly making a massive mistake: I was seconds away from pasting an OpenAI key into OpenClaw while debugging a failing request. I stopped myself, but barely.

When you paste an API key into an AI chat to debug an error, that key is permanently logged on someone else's server. Bots can scrape leaked keys within hours. Code scanners like GitGuardian and TruffleHog are great, but they only catch secrets in your repos. Traditional DLP tools completely ignore the AI chat box, which is exactly where you are typing.

I built a lightweight browser extension that catches secrets before they leave your machine.

A content script monitors 40+ AI chat sites. It intercepts the "Enter" key and the send button.

It scans your input locally using regex and high-entropy detection.

If a secret is found, you get a quick modal: Cancel, Send Anyway, or Redact. Messages without secrets pass through untouched. No slowdowns, no false interruptions.

Because a security tool shouldn't be a security risk, detection runs entirely locally. Your message content and secrets are never transmitted. The extension makes zero network requests in the free personal mode, aside from a daily signed config fetch for updated site lists. Detection rules are bundled locally.

Pricing & Tiers:

Free (Personal): API key detection on 40+ sites, no monthly cap, no signup required.
Pro ($7.99/mo or $59/yr): Adds PII detection (for 30 countries), custom regex patterns, one-click redaction, and warning history.
Team Tier: Force-install via Chrome Enterprise Policy, dashboard monitoring, and automated alerts to Slack, Teams, Telegram, Webhooks, or Email.

Links & Demos:

Try the interactive demo (no install needed): https://dariocositore.com/tools/api-key-leak-intercept/demo/
Chrome Web Store: https://chromewebstore.google.com/detail/api-key-leak-intercept/ajnlfcdamhbkpdkkmooijamanhlfhpod
Open source (MIT): https://github.com/dario-cositore/api-key-leak-intercept

I'll be hanging out in the comments all day! I'd love to answer any questions about the interception architecture, the redaction write-back problem, or anything else you're curious about.

All feedback, good or bad, is incredibly welcome.

Thank you all.

Dario