Tynapse

Tynapse builds an "AI Trust Layer" — runtime security infrastructure that sits between an enterprise application and its AI models, inspecting every response and tool call before it reaches a customer or executes an action. A two-stage Detect Guard and Trust Judge architecture blocks prompt injection, jailbreaks, PII leakage, hallucinations and unauthorized agent actions, while Trust Ops logs each decision as legally valid audit evidence for the Korean AI Basic Act, the EU AI Act, ISO 42001 and NIST AI RMF. Founded by former financial-sector CTO Minseung Kang with AI researchers from Google Research, the company is co-developing a runtime guardrail validation framework with KB Kookmin Bank and is expanding into healthcare, public sector and enterprise markets.