Tynapse

Tynapse is a Seoul, South Korea startup founded in 2025 that builds what it calls an “AI Trust Layer” – a runtime security layer that sits alongside deployed AI agents to detect and block risky behavior such as hallucinations, sensitive data leakage, and jailbreak attempts. The system uses a two-step detection-and-judgment structure and logs its decisions as audit records intended to hold up as legally usable evidence, a feature aimed squarely at regulated industries.

The company closed a seed round of roughly KRW 4.5 billion (reported as approximately $3.2-3.9 million depending on the source and exchange rate used) in April 2026, led by Mirae Asset Venture Investment with participation from Murex Partners and Kakao Ventures, following an earlier angel round and a pre-seed from Mashup Ventures. Tynapse is currently running proof-of-concept pilots with major Korean commercial banks and says it plans to expand into healthcare and public-sector use cases. It was a finalist, as the only Asian entrant, in both the NVIDIA GTC 2026 AWS Startup Pitch and the 2026 Snowflake Startup Challenge.

With roughly 15 employees and a product still in PoC stage with its anchor customers, Tynapse is an early but credible entrant in the fast-growing AI agent security space, led by CEO Minseung Kang, a former financial-sector CTO. Its claims about detection accuracy and audit-grade logging have not yet been independently tested or reported outside the company’s own and investor-sourced coverage.