Kyvvu
Kyvvu makes an in-process security kernel that checks each AI agent action against its task path before execution.
Founded
2025Kyvvu makes an in-process security kernel that checks each AI agent action against its task path before execution.
kyvvu.comCrawled Oct 16, 2026
The page explains how to report security vulnerabilities in Kyvvu's products or systems, what is in scope, and what to expect.
Kyvvu provides one set of rules for AI coding agents across machines.
The stack includes a private model hosted in Europe, an extensible agent harness, and a sandboxed runtime.
The page describes a hypothetical 2030 in which AI agents handle the noise while people focus on strategy.
The white paper explains what AI agents are, their security risks, and ways to secure them.
The company maps EU AI Act requirements to runtime needs for AI agents and shares its findings.
The company is building a runtime governance layer for autonomous AI agents.
The Agent Security Kernel for AI agents runs in-process, reasons over the full task, and keeps data in place.
An Agent Security Kernel (ASK) checks AI agents' actions against the full task path and allows, warns, or blocks them in-process.
How Kyvvu handles personal data in compliance with GDPR
Acceptable Use Policy for the Kyvvu platform
Terms of service for Kyvvu website and self-service account usage
GitHubRepository owner
Kyvvu introduces kyvvu-claude, a beta that checks Claude Code tool calls against policy before execution. It uses Claude Code’s own hooks to enforce policies, and it currently works with Claude Code.
Sep 2026 · KyvvuNews article
Kyvvu raises a €1 million pre-seed round from Volta Ventures and the Brabant Development Agency. The company says it will use the funding for its team and partner network.
Aug 2026 · KyvvuNews article
Dutch AI startup Kyvvu raises €1 million in seed funding to address a security gap in companies deploying AI agents.
Aug 2026 · JADSNews article