SecureAI PolicyGuard is an AI-driven enterprise data protection platform combining
- automated classification,
- policy-based encryption,
- audit-grade compliance logging.
- It bridges security operations, compliance, and IT automation with a dual-layer Desktop + REST API architecture.
| Capability | Description |
|---|---|
| AI Classification Engine | Detects sensitive content and calculates contextual risk scores |
| Policy Mapping Engine | Maps classification levels to automatic actions (Encrypt / Log / Ignore) |
| Encryption Layer (AES-256) | Local, per-directory key management using Fernet for authenticated encryption |
| Audit Logging | SQLite-based, immutable record of classification and encryption events |
| REST API (FastAPI + OAS 3.1) | Standardized interface for automation and remote analysis |
| Multi-Language GUI (DE/EN) | Dual-language interface for global operations |
| Batch Automation Script | One-click startup linking API server and GUI seamlessly |
+---------------------------------------------------+
| SecureAI PolicyGuard β Desktop GUI |
| - Configuration UI |
| - File scanning and visualization |
+---------------------------------------------------+
|
v
+---------------------------------------------------+
| Classifier Engine |
| - Heuristic + ML entity detection |
| - Risk calculation |
+---------------------------------------------------+
|
v
+---------------------------------------------------+
| Policy Engine |
| - Confidentiality mapping |
| - Action rules |
+---------------------------------------------------+
|
v
+---------------------------------------------------+
| Crypto Engine (Fernet / AES-256) |
| - Local encryption per workspace key |
| - Secure key rotation and storage |
+---------------------------------------------------+
|
v
+---------------------------------------------------+
| Audit Database (SQLite) |
| - Risk score, classification, entities, actions |
+---------------------------------------------------+
|
+--> REST API (FastAPI / OAS 3.1)
- Windows 10/11 (64-bit)
- Python 3.10+ (Recommended: 3.12)
- Internet access for dependency installation
pip install -r requirements.txtpython main.pystart.batAPI Endpoints:
POST /analyze-text
Content-Type: application/json
{
"text": "Customer IBAN: DE11 2222 3333 4444 00"
}{
"classification": "Highly_Confidential",
"risk_score": 88,
"entities": [["iban", "DE44...31"]]
}| Classification | Risk Range | Action |
|---|---|---|
| PUBLIC | 0β29 | Ignore |
| INTERNAL | 30β59 | Log Only |
| CONFIDENTIAL | 60β79 | Encrypt |
| HIGHLY_CONFIDENTIAL | 80β100 | Encrypt + Flag |
| Field | Description |
|---|---|
| file_path | Path of analyzed file |
| classification | Detected sensitivity level |
| risk_score | Calculated risk score |
| entity_count | Number of detected entities |
| timestamp | Timestamp of operation |
Stored in local SQLite database per workspace.
- Security Operation Centers (SOC)
- Compliance & Audit Departments
- Regulated industries (Finance, Energy, Government)
- Data Protection & Forensics Teams
SecureAI PolicyGuard includes automated Schemathesis-based API certification tests verifying OpenAPI 3.1 contract integrity.
- All test results are stored under
/tests/logs/for compliance evidence.
Developed by Thorsten Bylicki
- BYLICKILABS β Intelligence Systems & Communications
Β© 2025 Thorsten Bylicki | BYLICKILABS




















