OpenAI set to buy AI security startup Promptfoo to safeguard AI agents

OpenAI is preparing to acquire Promptfoo, a company known for helping enterprises detect and resolve security vulnerabilities in artificial intelligence models during the development phase. The move is part of OpenAI’s broader strategy to strengthen the safety and reliability of AI systems used by businesses.

The developer of ChatGPT is expected to announce that Promptfoo’s technology will be incorporated into OpenAI Frontier, a recently introduced platform designed to help organizations build, deploy, and manage AI agents. Frontier aims to ensure that these agents—often referred to by OpenAI as “AI coworkers”—operate within defined safety boundaries and have controlled access to data. Financial details of the acquisition have not been revealed.

OpenAI and other AI developers are competing to build more advanced AI agents capable of handling complex tasks with minimal human involvement. At the same time, companies are working to convince enterprises to adopt AI tools by emphasizing improved safety, efficiency, and reliability.

According to Ganesh Bell, the acquisition highlights OpenAI’s intention to make enterprise AI both powerful and secure at scale. Bell’s firm, Insight Partners, previously led an $18.4 million funding round for Promptfoo, with participation from Andreessen Horowitz.

Founded two years ago, Promptfoo develops open-source tools that allow organizations to test the security of AI systems and identify potential weaknesses. Its platform also enables companies to simulate cyberattacks on their own AI products—a process known as red-teaming—to uncover vulnerabilities before deployment. The San Francisco-based startup already serves roughly a quarter of Fortune 500 companies.

Promptfoo is part of a growing group of cybersecurity startups focused on using AI to defend against increasingly sophisticated cyber threats, even as malicious actors begin using AI to identify weaknesses in digital systems. OpenAI has also been strengthening the security capabilities of its own products. Recently, the company introduced an AI-powered tool designed to assist security teams in detecting and fixing vulnerabilities in large databases, similar to technology developed by competitor Anthropic.

With this acquisition, OpenAI plans to add automated security testing and red-teaming features to the Frontier platform. These capabilities will help organizations track testing activities, monitor system updates, and maintain compliance with security and risk management requirements. OpenAI also stated that it will continue supporting Promptfoo’s open-source projects while integrating its technology into the Frontier ecosystem.

Voice Of Osiz

The acquisition of Promptfoo by OpenAI highlights the growing importance of security in the rapidly evolving AI ecosystem. As businesses increasingly deploy AI agents to automate complex operations, ensuring safe and reliable AI environments has become a top priority. This move demonstrates how advanced security testing, red-teaming, and risk monitoring are becoming essential components of enterprise AI platforms. At Osiz, we view this development as a strong indicator that the future of AI innovation will be closely tied to robust security frameworks and responsible AI deployment. Organizations looking to adopt AI must integrate security from the development stage to prevent vulnerabilities and maintain trust. With the rise of autonomous AI systems, businesses will increasingly demand secure, scalable, and compliant AI infrastructures. This trend opens new opportunities for AI development companies to build intelligent solutions that combine innovation with strong cybersecurity practices.

Source: The Economic Times