What does Promptfoo sell?

Promptfoo sells Red Teaming, Guardrails, Model Security, MCP Proxy, Code Scanning and Evaluations. Red Teaming: Automated testing that generates attacks such as prompt injections and jailbreaks to find application-specific vulnerabilities in AI applications. Guardrails: Real-time filtering that protects AI applications against jailbreaks and adversarial attacks. Model Security: Security testing and monitoring for AI models. MCP Proxy: A proxy that secures Model Context Protocol communications.

How does CISO Tribune file Promptfoo?

CISO Tribune files Promptfoo under AI model and agent security, with AI firewalls and guardrails, Static application security testing and Breach simulation and automated pen-testing as secondary categories.

Where is Promptfoo headquartered?

Promptfoo is headquartered in San Francisco, United States.

Company record · security vendor

Promptfoo

Where Promptfoo appears on the record (4)

Every page on CISO Tribune and its network that carries this company, 4 in all.

In the vendor directory (4): AI model and agent security vendors · primary category
AI firewalls and guardrails vendors · category
Static application security testing vendors · category
Breach simulation and automated pen-testing vendors · category

What the company sells

Promptfoo sells AI security solutions, including an open-source pentesting and evaluation framework, that help developers and enterprises test and secure AI applications. It is based in San Francisco, California, was founded in 2024, and is now part of OpenAI while continuing as an open-source AI security project. Its product line covers automated red teaming for agents and RAGs, guardrails against jailbreaks, model security testing, an MCP proxy, code scanning for LLM vulnerabilities, and prompt evaluation tooling.^[2]

Categories: AI model and agent security (primary category)
AI firewalls and guardrails
Static application security testing
Breach simulation and automated pen-testing
Compare Promptfoo with companies in the same category
Headquarters: San Francisco, United States^[1]
Founded: 2024^[1]
Ownership: Owned by OpenAI
CISO on record: Not yet on our record. Send a verified source
Website: promptfoo.dev

Products and services

Product names as the company writes them; what each does, in our words, from the company's own product pages. Never how well it does it.

Red Teaming^[2]
Automated testing that generates attacks such as prompt injections and jailbreaks to find application-specific vulnerabilities in AI applications.
Guardrails^[2]
Real-time filtering that protects AI applications against jailbreaks and adversarial attacks.
Model Security^[2]
Security testing and monitoring for AI models.
MCP Proxy^[2]
A proxy that secures Model Context Protocol communications.
Code Scanning^[2]
Scanning that finds LLM vulnerabilities in a developer's IDE and CI/CD pipeline.
Evaluations^[2]
A framework for testing and evaluating prompts, models and RAG pipelines.

Competitors and alternatives

The other companies on our record filed under the same categories, category by category: first those that lead with that category and share more of Promptfoo's categories, then alphabetical. A list from the record, never a ranking.

On ai model and agent security, Promptfoo competes with SplxAI, AIShield, Alice, Enkrypt AI, HiddenLayer, Mindgard, Noma Security, Pangea, Straiker, Asymptote Labs and 45 more.
On ai firewalls and guardrails, Promptfoo competes with Impart Security, Lakera, Lasso Security, Operant AI, Prompt Security, SplxAI, AIShield, Alice, Enkrypt AI, HiddenLayer and 7 more.
On static application security testing, Promptfoo competes with AISLE, Checkmarx, GitHub Advanced Security, GitLab, Kiuwan, OpenText, Semgrep, Veracode, SplxAI, Fleuret AI and 11 more.
On breach simulation and automated pen-testing, Promptfoo competes with Fleuret AI, Armadin, Assail, Cymulate, Horizon3.ai, Pentera, Picus Security, SafeBreach, AttackIQ, Filigran and 4 more.

References

Every fact above cites one of these by number. Tier A is the company itself, a regulator or a filing; tier B is trade or business press.

[1]PressPromptfoo · promptfoo.dev · undated page, accessed November 4, 2026 · Tier A: the company, a regulator or a filing
[2]promptfoo.devPromptfoo · promptfoo.dev · undated page, accessed November 4, 2026 · Tier A: the company, a regulator or a filing

Last verified November 4, 2026. Our code found the quoted passage behind each fact on its source page when the fact was recorded, and re-checks every record against its sources each month.

Recorded November 4, 2026 from the references above.

Work for Promptfoo? Suggest a correction to a fact on this page (with a source on your own domain), or claim this entry by publishing a small file on your domain.

Security leadership

CISO Tribune has not yet recorded who holds the top security seat at Promptfoo. We add a seat only from a verified source; if you have one, send it to the desk.

Questions CISOs ask about Promptfoo

What does Promptfoo sell?: Promptfoo sells Red Teaming, Guardrails, Model Security, MCP Proxy, Code Scanning and Evaluations. Red Teaming: Automated testing that generates attacks such as prompt injections and jailbreaks to find application-specific vulnerabilities in AI applications. Guardrails: Real-time filtering that protects AI applications against jailbreaks and adversarial attacks. Model Security: Security testing and monitoring for AI models. MCP Proxy: A proxy that secures Model Context Protocol communications.
How does CISO Tribune file Promptfoo?: CISO Tribune files Promptfoo under AI model and agent security, with AI firewalls and guardrails, Static application security testing and Breach simulation and automated pen-testing as secondary categories.
Where is Promptfoo headquartered?: Promptfoo is headquartered in San Francisco, United States.
When was Promptfoo founded?: Promptfoo was founded in 2024.
Who owns Promptfoo?: Promptfoo is owned by OpenAI.

Related guides

CISO Reporting Structure, the Five Models and What the Filings Show
Most security chiefs report into technology leadership (64%), the rest to business leaders. The five models, their trade-offs and what 10-K filings show.
CISO News: Tracking Every Appointment and Departure in the Top Security Seat
CISO Tribune records every CISO appointment and departure as a sourced, dated entry, not a column: 466 moves to date, each with a primary source and a confirmed-or-reported label.
CISO, CSO, VP of Security, Head of Security: What the Titles Mean and What the Record Shows
A chief information security officer (CISO) owns information security. A chief security officer (CSO) usually adds physical security and investigations.