OpenAI to acquire Promptfoo

Source: OpenAI Date: 2026-03-09 URL: https://openai.com/index/openai-to-acquire-promptfoo

Summary

OpenAI announces the acquisition of Promptfoo — an open-source LLM testing and red-teaming framework widely used by developers to evaluate prompt performance, detect regressions, and test AI application safety. Promptfoo allows developers to run systematic evaluations against their AI pipelines. The acquisition brings prompt evaluation and testing tooling in-house for OpenAI, and likely signals intent to productize enterprise evaluation infrastructure.

Implications

The evaluation infrastructure thread. Promptfoo is a developer-beloved tool for prompt testing — acquiring it gives OpenAI the most-used open-source evaluation framework and the developer community that built around it. This is both a talent acquisition (the Promptfoo team understands evaluation deeply) and an infrastructure acquisition (integrating Promptfoo’s evaluation capabilities into OpenAI’s developer platform creates a complete loop: build prompts → test prompts → deploy prompts).

Acquiring the guardrails. Promptfoo is specifically used for red-teaming AI applications — finding failure modes, safety gaps, and unintended behaviors before deployment. OpenAI acquiring it signals that evaluation and safety testing tooling is core to their platform vision, not a third-party concern. This creates a potential moat: if developers evaluate AI applications in OpenAI’s platform, OpenAI gets signal about where AI pipelines fail and can improve their models accordingly.