LLM Gateway
Provider-agnostic routing with BYOK. See /products/llm-gateway.
Product · Implementation
How Privian implements the AI Security Layer
A single in-process hop between your application and AI providers.
This page describes how Privian builds the masking, routing and rehydration pipeline. For the category definition and reference architecture, see the educational pillar.
Overview
What it is
The AI Security Layer is Privian's product framing for the masking, routing and rehydration layer that sits between your application and AI providers. For the full category definition, architecture and FAQs, see the AI Security Layer category page.
How it works
Inside the request flow
Step 1
Detect
Scan inbound prompts for personal and sensitive entities.
Step 2
Mask
Replace detected values with deterministic placeholders.
Step 3
Route
Forward the masked prompt using your BYOK credentials.
Step 4
Rehydrate
Restore originals in the response inside the gateway.
Composed of
What the layer is composed of
PII Masking
Detection and replacement of personal and secret entities.
Prompt Security
Privacy-focused controls on what leaves your network as a prompt.
Rehydration
In-gateway restore of original values in responses.
FAQ
Frequently asked questions
- What is the AI Security Layer?
- A thin network hop between your application and AI providers that detects, masks, routes and rehydrates — without storing prompts or responses.
- How is it different from an LLM gateway?
- An LLM gateway is just routing. The AI Security Layer adds privacy controls — masking and rehydration — around that routing layer.
- Does the provider see the original values?
- No. The provider only ever sees the masked prompt.
Plans & pricing
Pricing for Privian's AI Security Layer
Explore plans for teams building privacy-sensitive AI workflows. Privian is in beta — pricing and limits may change.