Privacy on the edgeAI on real data.
Zero compliance risk.

PII intercepted before it reaches any AI model. No BAA required. HIPAA & GDPR compliant by architecture.

Get started free Request compliance review →

No BAA with your AI provider GDPR Art. 44 — no transfer of personal data HIPAA · GDPR · PCI

How it works

One proxy.
Two destinations.

Sending a patient record or a legal brief to OpenAI triggers HIPAA and GDPR obligations that almost nobody handles correctly. Privedge makes those obligations disappear — by architecture, not by contract.

Inspect

Each prompt is scanned for PII before a single token is sent anywhere.

Route

Sensitive prompts go to the edge model. The data never leaves the node.

Return

Clean prompts hit OpenAI or Anthropic as usual, zero added latency.

Live · try it

Watch a prompt get routed.

Detection runs on every keystroke, exactly like the worker at the edge. Add an email or an SSN and watch the route flip.

prompt

7 matches · regex detector v1

ROUTED TO

Edge model

7 sensitive entities detected. This prompt stays on the local node. Nothing reaches a third-party API.

Medical term ×4ES DNIEmailPhone

res.json()

{
  "routed_to": "edge",
  "model": "@cf/meta/llama-3.2-1b",
  "pii_matches": 7,
  "latency_ms": 179
}

What we detect

Sensitive data, caught
before it leaves.

v1 ships a fast regex detector.
v2 adds a named-entity model on Workers AI.

US SSN

123-45-6789

ES DNI / NIF

53412987Z

IBAN

ES91 2100 0418 45••

[email protected]

Phone

+1 555 203 9981

Card number

4716 •••• •••• 9012

Medical terms

diagnosis · dosage

Custom rules

your own regex →

compliance

Built for compliance from request one.

HIPAA

Health Insurance Portability and Accountability Act

GDPR

General Data Protection Regulation

PCI DSS

Payment Card Industry Data Security Standard

SOC 2 Type II

Service Organization Control 2

NIS2

Network and Information Security Directive 2

Attorney-client

Legal professional privilege

Full compliance documentation

Drop-in SDK

Change one import.
Keep your whole stack.

@privedge/sdk mirrors the OpenAI client one-to-one. Swap the import, point at your worker, and routing happens transparently.

ESM + CJS: ships both module formats, fully typed.

Zero added latency: clean prompts pass straight through.

routed_to on every response: audit exactly where each request went.

Read the docs @privedge/sdk

import Privedge from '@privedge/sdk'

// Same surface as the OpenAI SDK — change one import.
const ai = new Privedge({
  apiKey:       'YOUR_OPENAI_KEY',
  workerUrl:    'https://privedge.workers.dev',
  pii_strategy: 'edge',  // 'edge' | 'anonymize'
  edge_model:   '@cf/meta/llama-3.2-3b-instruct',
})

const res = await ai.chat.completions.create({
  model:    'gpt-4o',
  messages: [{ role: 'user', content: prompt }],
})

// routed_to tells you where each request went
console.log(res.routed_to)    // 'edge' | 'cloud'
console.log(res.pii_matches)  // detected PII types + counts

Privacy on the edgeAI on real data.Zero compliance risk.

One proxy.Two destinations.