Why I built a PII redaction API that doesn't use AI (and costs only $5/mo)

I shipped an AI feature last month and needed to redact user PII before sending prompts to GPT.

Every "privacy" API I evaluated did the same thing: sent my data to another AI model.

That's not privacy. That's a data leak with extra steps.

So I built something different.

The problem with AI-powered PII detection

If you're using Claude or GPT to detect SSNs, credit cards, and emails... you're literally giving the AI all that sensitive data. Most "privacy" services do exactly this.

The irony was too much. So I went back to fundamentals.

What PII Firewall Edge does differently

Zero AI dependency. Pure regex + 30 checksum validators:

- Credit cards → Luhn algorithm (same as your bank uses)

- Indian Aadhaar → Verhoeff (catches single-digit errors)

- Brazilian CPF → Dual Mod 11

- IBAN → Mod 97

Zero storage. Runs on Cloudflare edge (300+ locations). Nothing is logged, stored, or retained. Ever.

152 PII types. SSN, 50+ country IDs, 20 API key formats, crypto wallets, healthcare IDs.

Why $5/mo?

Enterprise PII solutions cost $50K+/year. I wanted to make this accessible to indie devs, startups building in public, and anyone who doesn't want to become a data liability.

Pricing:

- Free: 500 req/mo (for testing)

- Pro: $5/mo for 5,000 req

Try it: [RapidAPI](https://rapidapi.com/image-zero-trust-security-labs/api/pii-firewall-edge)

I'd love feedback. What PII types should I add?