Using the Private AI Gateway for Enterprise-Grade Privacy¶

Overview¶

The Private AI Gateway is 4Geeks AI Studio’s enterprise-grade privacy layer. It ensures your proprietary code never leaves your secure perimeter to train public models, using Zero Data Retention policies and private routing to all LLM providers.

In this tutorial, you’ll learn:

How the Private AI Gateway works
How to configure privacy settings
How Zero Data Retention is enforced
How to audit data flow and compliance
How to set up custom data handling policies

How the Private AI Gateway Works¶

Your Code ──────► Private AI Gateway ──────► LLM Providers
                      │
                      ├── Encrypted tunnel to all LLM APIs
                      ├── Zero Data Retention enforcement
                      ├── Request sanitization
                      ├── Response validation
                      └── Complete audit logging

Key Privacy Features¶

Feature	Description
Zero Data Retention	Your code is never stored by LLM providers
Encrypted Transit	All data in transit is encrypted (TLS 1.3)
Request Sanitization	Sensitive data (secrets, keys) is automatically removed
Response Validation	Output is checked for data leakage before returning
Audit Logging	Complete log of all AI requests and responses
Private Routing	Direct, private connections to LLM providers

Step 1: Enable the Private AI Gateway¶

Go to your AI Studio project settings
Navigate to AI Factory → Private AI Gateway
Toggle “Enable Private Gateway”
Select your privacy level:
Standard: Zero Data Retention with all providers
Enhanced: Standard + request sanitization + response validation
Maximum: Enhanced + private network routing + custom data policies

Step 2: Configure Zero Data Retention¶

Provider Agreements¶

The Private AI Gateway enforces Zero Data Retention with all LLM providers:

Provider	Zero Data Retention	Verified
Claude 4.5 (Anthropic)	Yes — API data not used for training	✅
GPT-5 (OpenAI)	Yes — API data not used for training	✅
Gemini 3 Pro (Google)	Yes — API data not used for training	✅
DeepSeek	Yes — API data not used for training	✅

Configuration¶

Go to Private AI Gateway → Data Retention
Verify Zero Data Retention is enabled for all providers
Review provider privacy agreements
Set data retention period for audit logs (default: 90 days)

Step 3: Configure Request Sanitization¶

The gateway automatically detects and removes sensitive data from requests:

Detected Patterns¶

Pattern Type	Examples	Action
API Keys	`sk-xxx`, `AKIAxxx`, `ghp_xxx`	Replace with `[REDACTED]`
Passwords	Hardcoded passwords in code	Replace with `[REDACTED]`
Database URLs	`postgresql://user:pass@host`	Replace with `[REDACTED]`
JWT Tokens	`eyJhbGciOi...`	Replace with `[REDACTED]`
Email Addresses	`user@company.com`	Optional: replace or keep
IP Addresses	`192.168.1.1`	Optional: replace or keep

Custom Patterns¶

Add your own sensitive data patterns:

custom_patterns:
  - name: "Internal API URLs"
    pattern: "https://api\\.internal\\.company\\.com/.*"
    action: "redact"

  - name: "Employee IDs"
    pattern: "EMP-\\d{6}"
    action: "redact"

  - name: "Customer Names"
    pattern: "customer_name:\\s*\"[^\"]+\""
    action: "redact"

Step 4: Set Up Response Validation¶

The gateway validates AI responses before returning them to your repository:

Validation Rules¶

Rule	Description
No data leakage	Ensures AI doesn’t echo back sensitive data
No hallucinated secrets	Checks for patterns that look like credentials
Code integrity	Verifies generated code doesn’t contain malicious patterns
Compliance check	Ensures output meets your compliance requirements

Configuration¶

Go to Private AI Gateway → Response Validation
Enable validation rules
Set action for failed validation:
Block: Reject the response and retry
Warn: Allow but flag for review
Log: Allow but log the issue

Step 5: Audit Data Flow¶

Audit Dashboard¶

Go to Private AI Gateway → Audit Log
View all AI requests with:
Timestamp: When the request was made
Model: Which LLM was used
Task: Associated AI task
Data sanitized: Whether sensitive data was detected and removed
Validation result: Whether the response passed validation
Retention status: Confirmation of Zero Data Retention

Export Audit Reports¶

Click “Export Report”
Select date range
Choose format (CSV, PDF, JSON)
Download for compliance documentation

Step 6: Configure Custom Data Policies¶

For enterprise customers with specific compliance requirements:

Data Residency¶

Specify where data can be processed:

Region	Available	Notes
US	✅	All providers have US endpoints
EU	✅	GDPR-compliant endpoints
LATAM	✅	Regional endpoints for Costa Rica, Brazil
Asia Pacific	Partial	Limited provider support

Compliance Frameworks¶

Map your gateway configuration to compliance frameworks:

Framework	Requirements	Gateway Support
SOC 2 Type II	Data protection, access controls	✅ Full support
HIPAA	Protected health information	✅ With BAA
GDPR	EU data protection	✅ Full support
PCI DSS	Payment card data	✅ Full support
ISO 27001	Information security management	✅ Full support

Best Practices¶

Privacy-First Development¶

Never commit secrets to your repository — use environment variables
Review sanitization logs regularly to catch new patterns
Update custom patterns when new types of sensitive data emerge
Audit quarterly to ensure compliance with your policies

Team Training¶

Educate developers on what constitutes sensitive data
Establish clear policies for handling credentials and secrets
Use secret management tools (Vault, AWS Secrets Manager) instead of hardcoded values
Review audit logs as part of your regular development process

Compliance Documentation¶

Export audit reports for compliance audits
Maintain data processing records as required by GDPR
Document your privacy configuration for internal reviews
Update policies when regulations change

What’s Next?¶

Learn about Getting Started with AI Studio
Explore Understanding the AI Factory
Read about Monitoring Token Usage

Need Help?¶

Documentation: docs.4geeks.io
Community: community.4geeks.io
Support: Available through the console dashboard

Still questions? Ask the community.