OpenAI Acquires Cybersecurity Startup Promptfoo to Strengthen AI Agent Security

OpenAI is moving to strengthen the safety of its AI tools by acquiring AI security startup Promptfoo, a company known for helping businesses test and secure artificial intelligence systems before they go live.

The deal, announced Monday, will bring Promptfoo’s technology into OpenAI’s enterprise platform, OpenAI Frontier. OpenAI said enterprises adopting AI agents need reliable ways to evaluate their behavior and detect potential risks before deployment.

The company plans to integrate Promptfoo’s capabilities into Frontier so organizations can automatically test for vulnerabilities such as prompt injections, jailbreak attempts, data leaks, and misuse of connected tools.

“Promptfoo brings deep engineering expertise in evaluating, securing, and testing AI systems at enterprise scale,” Srinivas Narayanan, CTO of B2B Applications at OpenAI, said in an announcement. “Their work helps businesses deploy secure and reliable AI applications, and we’re excited to bring these capabilities directly into Frontier.”

According to OpenAI, the platform will also introduce stronger oversight tools that allow companies to track testing activity, monitor changes over time, and maintain records to meet compliance and governance requirements.

Promptfoo’s tools already used by major companies

Founded just two years ago, Promptfoo has developed a suite of security and evaluation tools that help developers test AI applications and identify weaknesses before they reach production systems.

The San Francisco-based startup also created a widely used open-source command-line tool that allows developers to compare and test the performance of large language models.

OpenAI said the Promptfoo team, led by co-founder and CEO Ian Webster and co-founder Michael D’Angelo, will continue supporting the open-source project while also working on enterprise features integrated within Frontier.

Promptfoo’s technology is already widely adopted across large corporations. According to OpenAI, the company’s tools are trusted by more than 25% of Fortune 500 companies.

Promptfoo’s leadership said that joining OpenAI will allow the startup to scale its mission to help developers secure AI systems that increasingly interact with real-world data and tools.

“We started Promptfoo because developers needed a practical way to secure AI systems. As AI agents become more connected to real data and systems, securing and validating them is more challenging and important than ever,” Ian Webster, Co-founder and CEO at Promptfoo, said in the announcement. “Joining OpenAI lets us accelerate this work, bringing stronger security, safety, and governance capabilities to the teams building real-world AI systems.” 

The Promptfoo deal fits into a broader pattern of OpenAI using acquisitions to fill product gaps quickly. Earlier this year, the company acquired healthcare tech startup Torch for around $100 million in equity. Before that, it bought Software Applications, the maker of a Mac-based AI interface called Sky, and more recently hired Peter Steinberger, creator of a popular developer tool for building AI agents.

Also read: Agentic AI is already reshaping enterprise software, with platforms racing to embed autonomous tools into coding, search, and business workflows.