Kontext raises $4 million to police what AI agents do inside businesses
The Munich startup checks an agent's identity, requested resource and assigned task before allowing an action to run.
TLDR
Munich security startup Kontext has raised $4 million for software that enforces action-level policies on AI agents operating inside businesses. Its system sits between an agent and company tools, evaluating the agent's identity, the resource it wants and the task it was assigned before allowing or denying an action. Teams can begin in observe mode, then switch on enforcement with an audit log of decisions. 42CAP led the round, with a16z CSX and High-Tech Gründerfonds participating. Kontext says its product already supports coding agents including Claude Code and Codex.
Combined views
2.4K
4 Sources, first seen ago
