OpenAI to acquire Promptfoo to strengthen AI agent security testing

OpenAI says it plans to acquire Promptfoo, an AI security platform focused on finding and fixing vulnerabilities in AI systems during development. Once the deal closes, OpenAI says it will integrate Promptfoo’s technology into OpenAI Frontier, its platform for building and operating AI coworkers.

The move shows where enterprise AI is heading. As more companies deploy agents that can access tools, business data, and internal workflows, security testing is becoming a core part of product rollout rather than an optional extra. OpenAI said the goal is to strengthen agentic security testing and evaluation capabilities inside Frontier.

Promptfoo confirmed the deal in its own announcement and said the open-source project will remain open source. The company also said it will continue serving users and customers, which matters for developers who already use its tooling for red teaming and LLM evaluation.

OpenAI’s announcement says Promptfoo helps enterprises identify and remediate vulnerabilities in AI systems during development. The company said the integration will focus on native security testing and red teaming, workflow integration so teams can catch risks earlier, and stronger oversight and compliance features for governance and reporting.

That lines up with Promptfoo’s existing position in the market. OpenAI said Promptfoo is trusted by more than 25% of the Fortune 500 and described its open-source CLI and library as widely used for evaluating and red teaming LLM applications.

OpenAI also published an executive quote that explains why it wants the company. Srinivas Narayanan, CTO of B2B Applications at OpenAI, said Promptfoo brings “deep engineering expertise in evaluating and testing AI systems at enterprise scale,” adding that this will help businesses “deploy reliable applications with confidence.”

Promptfoo’s leadership framed the deal around growing demand for practical AI security controls. In its statement, the company said securing and validating AI agents is getting harder as these systems become more deeply connected to real-world data and workflows, and joining OpenAI should help it move faster on that work.

What OpenAI says the deal will add

Why this matters for enterprise AI

AI agents create a different security problem from standard chatbots. They can call tools, read internal data, and take actions across connected systems, which means failures can have a more direct business impact. OpenAI’s announcement explicitly mentions prompt injections, data leaks, unauthorized tool misuse, jailbreaks, and out-of-policy behavior as risks it wants to help enterprises catch earlier.

This acquisition also suggests OpenAI wants Frontier to compete on trust, not just capability. TechCrunch reported that the acquisition is meant to secure OpenAI’s AI agents, while OpenAI’s own post emphasizes testing, monitoring, compliance, and enterprise readiness. That points to a broader shift where agent platforms need strong security tooling built in from the start. This last point is an inference based on the official announcement and contemporaneous coverage.

Key details

• OpenAI announced the planned acquisition on March 9, 2026.
• Promptfoo separately confirmed it agreed to be acquired by OpenAI.
• OpenAI says Promptfoo’s technology will be integrated into OpenAI Frontier after closing.
• Promptfoo says its open-source project will remain open source.
• OpenAI did not disclose financial terms in its announcement.

FAQ