German security startup Kontext has secured $4 million in funding to develop software designed to manage the actions of AI agents within corporate environments. The round was led by 42CAP, with participation from a16z CSX and High-Tech Gründerfonds Management GmbH. The company intends to utilize the capital to grow its engineering team.
The platform functions as an intermediary between AI agents and corporate systems. It evaluates every requested action against established security policies, considering both the agent's identity and the specific task assigned. This task-aware approach prevents agents from performing unauthorized operations, such as modifying unrelated infrastructure or sending data to external services, even if the agent possesses the necessary credentials.
Kontext offers an observation mode that allows teams to test policies before enforcement. Once active, the system blocks unauthorized actions and maintains an auditable log of all decisions. The software currently supports coding agents such as Anthropic’s Claude Code and OpenAI’s Codex.
The startup was founded by Jens Ernstberger and Michel Osswald, both of whom hold doctorates in cryptography and computer security from the Technical University of Munich. The product is available to individual developers for free, while paid team plans start at $149 per month.
