The scariest button in tech is still “Send.”
Hey PH 👋
I’m Anuj — Cyber Security Professional, currently building OmniShield.
Quick context: I’ve watched Shadow AI get worse for years. Devs (and honestly, everyone) use ChatGPT/Claude for debugging, writing, life admin… and then paste keys, client data, bank stuff, medical notes without thinking. Blocking AI isn’t realistic. Ignoring the leak risk isn’t either.
So I built a Chrome extension that masks secrets at send time — on-device — on ChatGPT, Claude, Gemini, Cursor, etc. Free core. No “upload your prompt to us.”
Curious where you all land:
1) Have you ever almost (or fully) pasted something you instantly regretted into an AI tool?
2) Would you rather get a warning… or have it auto-masked before send?
3) For everyday users (not just company IT) — what’s the biggest trust blocker for installing a privacy extension?
Launching soon on PH. Would love brutal feedback before we go live:
Replies