promptShield - Anonymize client docs before you paste them into AI

Hi Product Hunt 👋 I'm a consultant, and promptShield started as a personal itch. Here's the thing nobody says (out loud): we're all using AI on client work - and we're all quietly downplaying it. For two reasons. One, we have no clean answer for the question "did you just paste my confidential file into ChatGPT?" Two - and this one runs deeper - the old world says an analyst is someone who writes the report. So if the AI wrote it, what are you even for? Admitting AI's role feels like admitting you're a fraud. But that definition is already dead. In the age of AI, an analyst isn't someone who writes the report - he's someone who vouches for someone else's report, someone else's solution. That's not a smaller job. It's a harder one (just ask managers). It takes more judgment, not less, to stand behind work and own the result in front of a client. The competence didn't go away - it moved. Once you see it that way, the shame disappears. You're not hiding that you used AI. You're doing the actual job: vouching. That leaves one real problem - the confidential data. And that one's just… solvable. So I built it. promptShield strips the names, emails, IBANs and addresses out of a document before anything reaches an AI, then decodes the AI's answer back on your machine so you get the real values. 100% offline - your documents never leave your laptop. No account needed to try it. The part that surprised me while building: every tool that already does this is aimed at enterprise dev teams - API keys, SDKs, procurement, cloud pipelines. There was nothing for the solo professional at a desk. So that's who this is for. Free for individuals. Stop hiding how you really work. Keep your edge, and keep your promise to your client. 🛡️ Would genuinely love your feedback - especially from other consultants, lawyers, and small firms.