Promptfoo
MIT-licensed LLM eval and red-teaming CLI, being acquired by OpenAI
Overview
Promptfoo is an open-source (MIT) CLI and library for testing, evaluating and red-teaming LLM apps, agents and RAG pipelines across dozens of model providers, with results wired into CI/CD. OpenAI announced in March 2026 that it would acquire the company; the Community edition stays free and open source.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- All LLM evaluation features
- All model providers and integrations
- Red teaming up to 10k probes a month
- Run locally or self-host
- Community support
- Custom red-teaming limits
- Team sharing and continuous monitoring
- Security and compliance dashboard
- SSO and granular permissions
- Managed cloud deployment; priority support and SLAs
- Deployment on your own infrastructure
- Complete data isolation
- Dedicated runner
- Assigned deployment engineer
What you can produce with Promptfoo
- Declarative YAML eval suites run from the CLI or CI/CD
- Dozens of model provider integrations, local models included
- Automated red teaming for prompt injection, jailbreaks, PII leaks and tool misuse
- GitHub Action that scans pull requests for LLM vulnerabilities
- LLM-as-a-judge and model-graded assertions
- Local or self-hosted runs with an MIT-licensed core
- Enterprise SSO, compliance dashboard, monitoring and on-premise deployment
Promptfoo is the eval and red-teaming tool many developers use instead of building their own harness: declarative YAML suites, dozens of providers, and a red-team mode that generates injection, jailbreak, PII-leak and tool-misuse attacks. The free Community edition covers most teams, with 10,000 red-team probes a month.
The headline is OpenAI's March 2026 agreement to acquire it on undisclosed terms. Both companies say the open-source project continues and the CLI is still MIT-licensed, but neutrality is now something to watch rather than assume. Enterprise and On-Premise are quote-only.
What we found
Promptfoo began as Ian Webster's open-source prompt-testing project, with a GitHub repository dating from April 2023, and became a company in 2024, co-founded with Michael D'Angelo and based in San Francisco. Andreessen Horowitz led a $5 million seed round in July 2024 and Insight Partners an $18.4 million Series A in July 2025.
On 9 March 2026 OpenAI announced it would acquire the company on undisclosed terms and fold Promptfoo's testing into OpenAI Frontier, its platform for building AI agents, while continuing the open-source project. Promptfoo's own post promised to keep supporting a diverse range of providers and models. Its About page now says Promptfoo is part of OpenAI, though neither company has published a dated closing notice that we could find. The CLI remains MIT-licensed and actively released; version 0.124.1 shipped on 8 October 2026.
The adoption figures are the vendor's own: more than 350,000 developers had used the tool and 130,000 were active each month when the deal was announced, and the homepage now claims 156 of the Fortune 500 as users. An earlier version of this page attached those numbers to the 2025 Series A, which reported 125,000 developers.
The work itself is declarative. A YAML file defines prompts, providers and assertions, and the same suite runs against OpenAI, Anthropic, Gemini, DeepSeek, Qwen, local models and dozens of other providers. Red-teaming mode generates attacks such as prompt injection, jailbreaks, PII leaks and insecure tool use, and a GitHub Action scans pull requests for LLM-specific vulnerabilities.
The free Community edition includes every evaluation feature and provider and up to 10,000 red-team probes a month, run locally or self-hosted. Enterprise and On-Premise are quote-only. Run locally, the CLI calls each model API directly, so token charges land on your own provider account. The open question is direction: a testing tool owned by a model lab now has to keep showing that it is neutral.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including Promptfoo's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Promptfoo unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Promptfoo directly →
Spotted something out of date? Suggest an update →
Related reading
More in Code & Dev Tools
