In a significant move to enhance the security and trustworthiness of artificial intelligence agents, OpenAI announced on Monday its acquisition of Promptfoo, a pioneering AI security startup. This strategic acquisition underscores the growing importance of securing AI systems, particularly as the development of independent AI agents capable of performing complex digital tasks accelerates.
Promptfoo, founded in 2024 by Ian Webster and Michael D’Angelo, has quickly established itself as a leader in developing tools to safeguard Large Language Models (LLMs) from various online threats and vulnerabilities. The company provides an open-source interface and library that allows enterprises to test and identify security weaknesses within their AI models during the development phase. Its impressive client roster reportedly includes over 25% of Fortune 500 companies, a testament to the critical need for robust AI security solutions.
The financial terms of the transaction were not disclosed, though Promptfoo had previously raised $23 million and was valued at $86 million after its most recent funding round in July 2025, according to PitchBook. This acquisition marks a clear intent from OpenAI to proactively address the security challenges inherent in advanced AI deployments.
Upon the deal's closure, Promptfoo’s technology will be seamlessly integrated into OpenAI Frontier, OpenAI’s dedicated enterprise platform for AI agents. This integration is poised to bring several key capabilities to OpenAI's ecosystem. Specifically, Promptfoo's tools will enable automated red-teaming exercises, allowing for systematic identification and mitigation of potential vulnerabilities within AI systems. Furthermore, the technology will facilitate the rigorous evaluation of agentic workflows for security concerns and provide continuous monitoring for risks and compliance requirements.
The rapid advancement of independent AI agents, which hold immense promise for boosting productivity across various sectors, also introduces new avenues for malicious actors. These agents, while powerful, could potentially be exploited to access sensitive data or manipulate automated systems if not adequately secured. OpenAI's move to acquire Promptfoo highlights a broader industry recognition that for AI agents to be truly transformative and widely adopted, their safety and reliability must be paramount.
OpenAI has also indicated its intention to continue developing Promptfoo's open-source offerings, reinforcing its commitment to fostering a secure and collaborative AI development environment. This dual approach – integrating advanced security features into its proprietary platforms while also contributing to the open-source community – is crucial for building trust and accelerating the safe deployment of AI technologies.
This acquisition sends a strong message to the industry: the era of AI agents is here, and with it comes an urgent need for advanced security protocols. By bringing Promptfoo's expertise in-house, OpenAI is not only strengthening its own offerings but also setting a precedent for how frontier AI labs should prioritize safety and resilience in the face of increasingly sophisticated cyber threats. As AI continues to evolve, investments in security will be as critical as advancements in capabilities, ensuring that these powerful tools can be harnessed responsibly for the benefit of all.
