📸 The Enterprise Studio in Action

Interactive PII Playground (Live Detection with Visual Highlight Pills)

Inspect, scrub, and redact sensitive entities with sub-25ms latency. Visual badge pills pinpoint exactly what the engine detected before it ever leaves your network.

⚡ Overview

Enterprise PII Guardrails Studio is an ultra-fast, defense-in-depth privacy gateway designed to intercept, detect, mask, and pseudonymize Personally Identifiable Information (PII) before it reaches third-party LLM APIs (OpenAI, Anthropic, Google Gemini, Azure OpenAI) or internal vector databases.

Key Highlights:

• Sub-25ms Live Masking: In-memory LRU authorization cache and non-blocking asynchronous audit pipeline.
• Zero-Leak Guarantee at Rest: Database, API keys, and audit logs are encrypted using SQLCipher AES-256-CBC.
• SHA-256 Cryptographic Verification: Every downloaded runtime engine is verified on the fly against official SHA-256 hashes prior to execution.
• Node-Locked Licensing (v4): Commercial licenses are cryptographically locked to the host's Server Hardware ID via Ed25519 asymmetric signatures—operating 100% offline with zero cloud dependency.
• Bi-Directional Token Restoration: Effortlessly restore LLM responses (/unmask) back to original values for authorized end users.

🔄 How It Works (Zero-Trust Pipeline)

sequenceDiagram
    autonumber
    actor User as Client / App
    participant GW as PII Guardrail Gateway
    participant Cache as In-Memory Auth Cache
    participant Engine as Multi-Tier PII Recognizer
    participant DB as SQLCipher AES-256
    participant LLM as External LLM (OpenAI / Anthropic)

    User->>GW: POST /mask (Prompt with sensitive PII)
    GW->>Cache: Verify X-API-Key (< 0.1ms)
    GW->>Engine: Scan regex rules + NLP entities
    Engine->>DB: Asynchronously store reversible tokens & mapping
    GW-->>User: Return Safe Payload (<PERSON_1>, <SSN_1>)
    
    User->>LLM: Send sanitized payload
    LLM-->>User: Return completion with tokens
    
    User->>GW: POST /unmask (LLM output + mapping vault)
    GW-->>User: Return restored response with real entity values

🚀 Quickstart

Method 1: Using pip (Recommended)

# 1. Install from PyPI
pip install piiguardrails

# 2. Launch the studio
piiguardrails

On first launch, piiguardrails automatically streams the core binary, verifies its SHA-256 cryptographic checksum, provisions your encrypted database, and opens the studio dashboard on: 👉 http://localhost:8000

Method 2: Using uv (Instant Sandbox — Zero Setup)

Run directly in a self-contained, isolated environment without polluting global Python:

uv run --with piiguardrails piiguardrails

Method 3: Standalone Pre-Compiled Executable

For restricted air-gapped environments without Python installed, download the standalone binary directly from GitHub Releases:

• Windows x64: EnterprisePIIGuardrail-v2.0.6-windows-x64.exe
  • SHA-256: 1f910b35635edb8ceebffe999513e77d037ca3a918724c1fb0ebc939da8a1e4e
• Linux x86_64: EnterprisePIIGuardrail-v2.0.6-linux-x86_64
  • SHA-256: 50c7f948d724082ad007eab30c7f53b391ed8a5249237e8e1bc3d535eceadc4f

Run directly:

# Linux
chmod +x EnterprisePIIGuardrail-v2.0.6-linux-x86_64
./EnterprisePIIGuardrail-v2.0.6-linux-x86_64

# Windows
.\EnterprisePIIGuardrail-v2.0.6-windows-x64.exe

Method 4: Docker / Podman Deployment

docker run -d \
  -p 8000:8000 \
  -v $(pwd)/data:/app/data \
  -e SQLCIPHER_PASSPHRASE="your-secure-passphrase-min-32-chars!" \
  --name pii-guardrail-studio \
  piiguardrails/enterprise-pii-guardrail:latest

🎁 Community Launch Promo: 6 Months Free Enterprise

To celebrate our v2.0 release, early adopters can unlock full Enterprise Tier features (unlimited API requests, unrestricted payload length, and custom regex policies) through March 31, 2027:

ED3-AMBGXK7UVD777777-GRYXEUYUZ432JOHY-UWVDBHXRLAXU4U47-7MBK2DQGIUV4JQT6-UJNRXNTHI3JPBGIS-P66HGKVYNLMQXHHS-M3N4RF3XDN6LZILT-WQNCGJ4KEY6ONIIM

To activate:

1. Open the studio dashboard (http://localhost:8000).
2. Go to Settings > License & Quota in the sidebar.
3. Paste the key above and click Activate License.

🧩 Python Integration Example

Once the studio is running, integrate it directly into your LangChain, LlamaIndex, or raw API pipeline:

import requests

API_KEY = "your-api-key-here"  # Generated in the Studio UI
headers = {"X-API-Key": API_KEY, "Content-Type": "application/json"}

# 1. Mask sensitive input before sending to LLM
payload = {
    "text": "Hello Alice Smith, your verification code was sent to alice.smith@acme.corp. Call us at 415-555-2671."
}

mask_response = requests.post("http://localhost:8000/mask", json=payload, headers=headers).json()
print("Masked text for LLM:", mask_response["masked_text"])
# Output: "Hello <PERSON_1>, your verification code was sent to <EMAIL_1>. Call us at <PHONE_1>."

# 2. Query your LLM with safe, masked prompt...
# simulated_llm_reply = "We contacted <PERSON_1> at <EMAIL_1> regarding the ticket."

# 3. Unmask the LLM response back to real entities for your user
unmask_payload = {
    "text": "We contacted <PERSON_1> at <EMAIL_1>.",
    "mapping": mask_response["mapping"]
}
unmask_response = requests.post("http://localhost:8000/unmask", json=unmask_payload, headers=headers).json()
print("Restored response:", unmask_response["unmasked_text"])
# Output: "We contacted Alice Smith at alice.smith@acme.corp."

🛡️ Enterprise Feature Matrix

📚 Complete Documentation Index

• 📖 API Reference Specification: Full OpenAPI / REST documentation with request/response schemas for /mask, /unmask, /health, and Admin endpoints.
• 🔑 Customer Licensing Guide: Guide on tiers, obtaining your Server Hardware ID, and activating enterprise keys offline.
• 💾 Backup & Disaster Recovery Guide: Database persistence, .env encryption passphrase management, and server migration.

📋 System Requirements

• Operating System: Windows 10/11/Server (64-bit) or Linux (Ubuntu, Debian, RHEL, CentOS, Rocky, WSL2 x86_64)
• Python: >= 3.8 (if running via pip / uv)
• Memory: Minimum 2 GB RAM (4 GB recommended for high throughput)
• Network Port: Default 8000 (configurable via --port or .env)

📄 License

This software is governed by the Enterprise Evaluation and Commercial License Agreement. All rights reserved.