OpenAI announced on March 9 that it is acquiring Promptfoo, a cybersecurity startup whose tools help developers test and secure large language models and AI agents. The deal, for which terms were not disclosed, will bring Promptfoo's entire team inside OpenAI and integrate its technology into the company's Frontier platform for AI coworkers.[1]
Promptfoo was founded in 2024 by Ian Webster and Michael D'Angelo to build red-teaming and evaluation tooling for LLMs. Despite raising just $23.6 million in total funding - including an $18.4 million Series A led by Insight Partners, with participation from existing investor Andreessen Horowitz, in July 2025 - its open-source library reached adoption at more than 25% of Fortune 500 companies, according to the company.[2] Its last private valuation stood at $86 million.[2]
The strategic logic is straightforward. As AI agents gain access to real enterprise data, external APIs, and automated workflows, the attack surface expands considerably. Prompt injection, data exfiltration via agentic chains, and compliance gaps in automated decision-making are not theoretical concerns - they are active problems for enterprises deploying agents today. By folding Promptfoo's capabilities into Frontier, OpenAI is betting that first-party security tooling will become a key differentiator in the enterprise market.
OpenAI said the integrated platform will support automated red-teaming, security evaluation of agentic workflows, and real-time risk and compliance monitoring.[1] The company also committed to continuing development of Promptfoo's open-source project - a notable pledge given that open-source credibility has been central to Promptfoo's developer adoption.[1]
"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, Promptfoo[1]
"Promptfoo brings deep engineering expertise in evaluating, securing, and testing AI systems at enterprise scale. Their work helps businesses deploy secure and reliable AI applications, and we're excited to bring these capabilities directly into Frontier." - Srinivas Narayanan, CTO of B2B Applications, OpenAI[1]
The acquisition fits a broader pattern. OpenAI acquired health-care tech startup Torch in January and Software Applications, maker of Sky, an AI interface for Mac that could see a user's screen and act across apps, in October 2025. In February, it hired Peter Steinberger, the developer whose open-source personal agent framework OpenClaw went viral earlier this year, accruing over 180,000 GitHub stars and prompting Sam Altman to describe Steinberger's new role as driving "the next generation of personal agents."[3] The company is assembling a vertical stack around its agent platform - infrastructure, interfaces, and now security - as competition with Anthropic, Google, and Meta intensifies.
Sign in to join the discussion.
What makes the Promptfoo deal notable is less its price tag than its signal: security is no longer an afterthought to be delegated to third-party vendors. For OpenAI, it is now a product feature.