OpenAI acquires Promptfoo to secure its AI agents

OpenAI announced the acquisition of Promptfoo, a security startup founded in 2024 by Ian Webster and Michael D’Angelo, to bolster the safety of its AI agent platform. The deal, for an undisclosed sum, brings in technology already trusted by more than 25% of Fortune 500 companies. Promptfoo, which had raised $23 million and was valued at $86 million, specializes in tools for testing security vulnerabilities in large language models (LLMs), including an open-source interface and library.

Once integrated, Promptfoo's capabilities will be a core part of OpenAI Frontier, the company's enterprise platform for autonomous AI agents. The technology will enable automated red-teaming—simulating adversarial attacks—to proactively find weaknesses. It will also evaluate complex, multi-step agentic workflows for security concerns and provide continuous monitoring for risks and compliance. This move directly addresses growing industry fears that autonomous agents performing digital tasks could be manipulated to access sensitive data or disrupt automated systems.

OpenAI stated it plans to continue developing Promptfoo's open-source offerings, suggesting a commitment to broader ecosystem security. The acquisition highlights the intense pressure on frontier AI labs to prove their most advanced systems can be deployed safely in critical business operations. As enterprises increasingly adopt agentic AI for productivity gains, robust, built-in security evaluation is becoming a non-negotiable requirement for platform adoption.