Your AI prompts are leaking personal information right now. Stop it.

⚠ Every prompt without DDS Protect is a compliance risk

The $229 self-hosted PII redaction sidecar that keeps regulated data on your servers — not your LLM provider's. One Docker container. One environment variable change. Zero data egress.

🔒 Stop the Leak — Buy Now $229
dds-protect — redaction proxy :8080
Trusted by engineers who build with
OpenAI Privacy Filter Apache 2.0 HuggingFace Docker
// The Problem

The Data You're Sending to OpenAI Right Now

Every prompt your app sends to an LLM API contains raw, unredacted user data. Names. Emails. Phone numbers. SSNs. Credit cards. Sent in plaintext. Sitting on third-party servers. Your compliance team probably doesn't know.

€0M GDPR Fines
0% Global Revenue at Risk
0ms Detection Latency
DDS Protect before and after redaction demonstration — raw PII on the left, redacted categories on the right
Live redaction: every PII field replaced with its category before reaching OpenAI
// Architecture

Five Minutes. One Container. Done.

DDS Protect sits between your app and any OpenAI-compatible LLM API. It detects and redacts PII before a single byte leaves your infrastructure.

📱
Your App
OpenAI SDK
→
🐳
DDS Protect
Docker · :8080
→
🔒
Redacted
[CATEGORY] labels
→
🤖
LLM API
OpenAI / Anthropic
DDS Protect architecture: Your App → DDS Protect proxy container → Redacted request → LLM API → Response
Architecture: DDS Protect sits as a transparent sidecar between your app and any LLM API
$ docker compose up -d # ↑ That's it. One command. Five minutes. $ export OPENAI_BASE_URL=http://localhost:8080/v1 $ # Your existing OpenAI SDK calls work unchanged.
DDS Protect Docker deployment — single container, one environment variable change
Deploy anywhere Docker runs — on-prem, cloud, or that dusty Dell PowerEdge
// Features

Everything You Need. Nothing You Don't.

Eight capabilities that turn your AI features from "compliance liability" to "auditor-ready" in five minutes.

🏠

Zero Data Egress

So what: No regulatory exposure. Period.

DDS Protect runs entirely on your hardware. Detection in your container's memory. Redaction inside your network. No telemetry. No license check phoning home. The only thing that leaves: sanitized text.

🎯

8 Built-In Categories · Fine-Tune to 50+

So what: Catch what auditors care about — then teach it what yours care about.

Out of the box: names, emails, phones, addresses, SSNs, credit cards, dates of birth, IPs — 96% F1 via OpenAI's open-source privacy-filter model (Apache 2.0). The model is fine-tunable: the community has already extended it from 8 to 50+ entity types using domain-specific training data. Medical record numbers, passport IDs, custom organizational codes — if you can label it, the model can learn it. Fine-tune once, deploy forever.

⚡

Transparent Drop-In

So what: Zero engineering distraction.

Change one environment variable: OPENAI_BASE_URL from api.openai.com to localhost:8080. Your existing SDK calls, LangChain pipelines, and custom clients work without modification. Five minutes from download to redaction.

🌊

SSE Streaming

So what: Your users notice nothing.

Server-Sent Events pass through transparently. Chat messages stream token-by-token exactly as they do today. 50–200ms detection latency happens before the first token — zero perceived delay for your users.

📜

Immutable Audit Trail

So what: Prove compliance in 60 seconds.

Every detected PII span logged to a local SQLite database: category, original value, position, confidence score, timestamp. Append-only — no record ever modified or deleted. This is your legal defense, not just a feature.

🖥️

Browser Dashboard

So what: CISO visibility without a terminal.

audit-viewer.html ships with every download. Open in any browser — connects to the sidecar's /audit API endpoint for live redaction data. Total events, category breakdowns, confidence distributions, searchable event log. Auto-refreshes; just point it at your running sidecar.

🐋

Runs Anywhere Docker Runs

So what: No infrastructure religion required.

Linux, macOS, Windows. On-prem, AWS EC2, that dusty Dell PowerEdge. CPU-only: 8GB RAM. GPU (CUDA / Apple Silicon MPS): auto-detected, drops latency from ~150ms to ~30ms.

📖

Apache 2.0 Model

So what: No vendor lock-in. No license fees.

The detection model — openai/privacy-filter — is Apache 2.0 licensed, publicly available on HuggingFace. Inspect the model card. Verify benchmarks independently. ~700 lines of readable Python wrapper. If we disappeared tomorrow, your deployment keeps running.

DDS Protect audit dashboard — view total events, category breakdowns, and searchable event log
Audit dashboard: open in any browser. No server. No credentials. Reads SQLite directly.
// Pricing

$229. Once. Yours.

Enterprise PII redaction gateways start at $30,000–$50,000 per year. If you're not Fortune 500, that price isn't expensive — it's disqualifying. But GDPR fines don't scale down for smaller companies.

GDPR fines reach €20M or 4% of global revenue
DDS Protect costs $229. Once.
🛡️ 30-Day Money-Back Guarantee 🔥 Launch Special — Limited Time $299$229 One-time payment · Perpetual license · Royalty-free commercial use vs $30,000–$50,000/year for enterprise alternatives
What you get ✓ Full source code (~700 lines Python)
✓ Docker image + docker-compose.yml
✓ audit-viewer.html dashboard
✓ GDPR/HIPAA DPIA compliance template
✓ Setup guide · 30 days email support
What you don't get ✗ Ongoing support beyond 30 days
✗ Hosted/SaaS version
✗ Automatic model updates
✗ Custom model fine-tuning
✗ Someone to deploy it for you

Detection model: openai/privacy-filter — Apache 2.0, free for commercial use, ~3GB, cached locally on first run.

🔒 Get DDS Protect — $229
🛡️ 30-Day Money-Back Guarantee — No questions asked. Full refund if it doesn't meet your needs.
// Trust

No Logos Yet. Here's What We Do Have — And Why It Matters More.

We're new. But what we ship is open, auditable, and built on battle-tested architectural patterns.

DDS Protect compliance: GDPR, HIPAA, SOC 2 Type II, CCPA ready
Compliance-ready: GDPR, HIPAA, and DPIA template included with every purchase
🔓

Open Model

openai/privacy-filter — Apache 2.0. Publicly available on HuggingFace. Download it, benchmark it, verify the 96% F1 claim yourself. No proprietary black box.

👁️

700 Lines of Python

The entire DDS Protect wrapper is ~700 lines of readable, documented Python. A senior engineer can read the full codebase in an afternoon. No hidden functionality. No telemetry. No surprises.

🏗️

Battle-Tested Pattern

Transparent sidecar proxy — the same architectural pattern used by Kong, Envoy, and Apigee. Applied to one specific, high-risk use case: preventing PII from reaching LLM APIs.

📋

Compliance Ready

Every purchase includes a 444-line DPIA template for GDPR Article 35 and HIPAA. Maps each PII category to specific regulations. Drop it into your existing DPIA with minimal modification.

DDS Protect trust: open-source model, auditable code, self-hosted deployment
Self-hosted. No telemetry. No license checks phoning home. Your data stays on your hardware.
// Compare

$229 vs $50K/yr vs €20M Fine

You have three choices. Two of them are expensive. One of them is five minutes.

✓ RECOMMENDED
Self-Hosted
DDS Protect
$229
One-time · Perpetual license
  • ✓ 8 PII categories · 96% F1
  • ✓ Self-hosted · Zero egress
  • ✓ Drop-in proxy · 5 min setup
  • ✓ Immutable audit trail
  • ✓ Docker · Runs anywhere
  • ✓ Apache 2.0 model
  • ✓ DPIA template included
  • ✓ 30-day money-back guarantee
Enterprise
API Gateway Vendor
$50K+
Per year · Per seat
  • ✓ PII detection (varies)
  • ? May phone home
  • ? Complex Integration
  • ✓ Audit logging
  • ? Vendor lock-in
  • ✗ Proprietary model
  • ✓ Compliance features
  • ✓ Enterprise support
⚠ HIGH RISK
Do Nothing
Status Quo
€20M
Potential GDPR fine · 4% revenue
  • ✗ Raw PII → OpenAI
  • ✗ Zero data protection
  • ✗ No audit trail
  • ✗ GDPR violation
  • ✗ HIPAA violation risk
  • ✗ No compliance docs
  • ✗ Data leaves your network
  • ✗ Career-ending breach
// After Purchase

Four Steps. Five Minutes. Done.

From Stripe receipt to PII redaction in production — here's exactly what happens after you buy.

01 📦

Download

Receive a download link instantly after purchase. Contains full source code (~700 lines), Docker image configs, and the dashboard. No waiting for account provisioning.

02 ⚙️

Configure

Set OPENAI_API_KEY and OPENAI_BASE_URL in your environment. That's it. The detection model downloads and caches automatically on first run.

03 🚀

Deploy

docker compose up -d — DDS Protect starts proxying requests on port 8080. Your existing OpenAI SDK calls work unchanged. No code changes in your app.

04 ✅

Verify

Open audit-viewer.html in your browser. Watch PII detections arrive in real-time. Check your DPIA template. Show your compliance team. You're done.

Every prompt without DDS Protect is a compliance risk 🔒 Get DDS Protect — $229 🛡️ 30-Day Money-Back