Mask PII locally before it reaches any LLM

How NoeticGuard works

An egress firewall for AI: text is tokenized before it leaves toward ChatGPT, Claude, Gemini, OpenAI-compatible APIs, or your own model host. Pick the surface that matches the job.

SDK path — your product → any LLM

Browser Shield runs this loop inside the extension. For your own app, wire the same four calls on your server:

1. 1. Policy sync — GET /v1/config with a publishable key.
2. 2. Local mask — maskPii() in-process before any outbound LLM call.
3. 3. Safe egress — Only tokens like [EMAIL_1] reach the model provider.
4. 4. Telemetry — Report masked_count via POST /v1/telemetry. maskPii() does not do this by itself — without this POST, Home usage stays at 0.

What it detects

High-precision detectors (regex + validators such as Luhn / IBAN / TC kimlik). Designed to avoid shredding normal prose.