Healthcare payers & providers
Kagetora — AI Security Testing for Regulated Healthcare
Kagetora is Karunyx's AI security testing platform for regulated healthcare: tiered prompt-injection and AI safety detection through a deterministic and semantic judge architecture, for systems operating near PHI.
Capabilities
Deterministic + semantic judges
Tiered detection combines deterministic checks with semantic judgment, so obvious attacks are caught cheaply and subtle ones still get a real evaluation.
Evaluation harness with real scoring
Precision, recall and F1 scoring on the detection layer itself — because "we have guardrails" isn't an answer a security reviewer can accept without numbers.
NeMo Guardrails integration
Works with the guardrail tooling teams are already standing up, rather than asking them to replace it.
PHI redaction and audit logging
PHI redaction is enforced pre-inference, and every evaluation produces structured audit logging built for the payer/provider security review process.
Technical controls
Kagetora is designed for HIPAA and HITECH obligations and a zero-trust security architecture for PHI. These are the frameworks the product is built to align to — not certifications it holds, and not audits it has completed.
- AES-256
- Encryption at rest
- TLS 1.3
- Encryption in transit
- IAM
- Identity and access management
- Multi-tenant isolation
- Tenant boundaries enforced by design
- PHI redaction
- Enforced pre-inference, before the model sees data
- Audit trails
- Structured logging built to be admissible
Where it stands
Under active development
Kagetora is under active development. We haven't been through a third-party compliance audit, and we won't imply otherwise — what we can show you is the architecture, the evaluation results, and the audit trail it produces.
See a sample assessment report →Where it comes from
Built from the payer side out
Kagetora's design choices come directly from Samuel Beera's experience building production AI infrastructure at a large healthcare payer organization. He now leads AI and agentic systems architecture at a large healthcare provider organization — which keeps both halves of this product's audience in view.
More about the founders →Put it in front of your security review
If you're the one who has to sign off on an AI system touching PHI, we'd like to show you what Kagetora produces.