GenCrafter Tracks AI Interactions Silently and Securely
AI Tracker inspects outbound LLM traffic at the network/proxy layer no browser plugins required. It analyses prompts and file endpoints, applies rule + model-based detection, and enforces policies in real time while generating audit-ready logs.
What GenCrafter Can Do
- Monitor without plugins runs at network/proxy layer
- Detect sensitive data (PII, PHI, credentials, source code)
- Enforce policies: block, alert, redact, or log
- Map AI actions to users via SSO, IP, or device
- Block or restrict file uploads to LLMs (e.g. large files, sensitive documents)
- Effortless audit ready for GDPR and HIPAA Compliance.
- Real‑time monitoring instantly tracks all LLM interactions for immediate risk detection.
How It Works
Traffic to public or private LLMs is routed through GenCrafter’s AI Gateway. Prompts and file uploads are scanned for PHI/PII patterns and contextual risks. The policy engine evaluates each action and either allows, redacts, alerts, or blocks it. All interactions are logged for audits and compliance.
Deployment Options
-
Cloud (Managed)
UAE/EU regions, auto-scaling, managed infra -
On-Prem
VM cluster with GPU nodes for local control -
Hybrid
Central ML cluster + lightweight proxies for remote clinics
Key Capabilities
- Regex + ML pattern detection (tuneable)
- Supports more than 60+ PII/PHI categoriesincluding personal, financial, healthcare, and digital identifiers
- Designed for integration into internal compliance systems or larger AI governance platforms
- Client-defined custom labels, categories, rules
- File endpoint interception & blocking
- Exportable compliance reports (CSV / PDF)
- SSO/IP/device mapping & audit trails
Use Cases
- Enterprise AI Governance visibility & controls without disrupting workflows.
- Data Loss Prevention stop PII/PHI/credentials from reaching external LLMs.
- Security & Risk Management real-time monitoring, alerts & forensics.
- Audit-Ready Logging compliance reporting for regulators and auditors.
No by default AI Tracker blocks file endpoint calls before they
reach external servers. With customer opt-in on-prem deployments we
can optionally inspect files locally for classification purposes.
We use pattern-based detection combined with ML models to identify
sensitive healthcare and personal data in real time.
Yes. The platform is designed to help organizations meet HIPAA,
GDPR, and EU AI Act requirements.
Yes supports regex/pattern detection and ML contextual detection. We
tune during PoC to match your identifier variants.
UAE/EU cloud regions, on-prem clusters, or hybrid depending on data
residency and compliance needs.
The PoC supports up to 1,500 concurrent users and includes full
setup, sizing, and an executive report. It is offered at low cost
and may be free for qualifying enterprises.
Production licensing is available as a self-hosted model with volume
discounts for organizations with 5,000+ users. Managed-cloud pricing
is also available upon request.
