Skip to main content

Secure your AI
from prompt to production

Open-source LLM security trusted by 75,000+ users

Red Teaming

Adaptive red teaming that targets applications, not just models

Generate customized attacks for your use case:

npx promptfoo@latest redteam init

Our language models probe for specific risks in your system.

  • Direct and indirect prompt injections
  • Jailbreaks tailored to your guardrails
  • Data and PII leaks
  • Insecure tool use vulnerabilities
  • Unauthorized contract creation
  • Toxic content generation
  • And much more

» Learn more about Red Teaming

Trusted by teams at

ShopifyDiscordAnthropicMicrosoftDoordashCarvana

Why Promptfoo?

Find vulnerabilities you actually care about

Our specialized language models generate attacks specific to your industry, company, and application.

No generic canned attacks - every attack is created on-the-fly.

Learn More
promptfoo security coverage examples

Battle-tested at enterprise scale

Adopted by 27 Fortune 500 companies shipping apps to hundreds of millions of users.

Embraced by an open-source community of over 75,000 developers worldwide.

Get Started
promptfoo quickstart

Security-first, developer-friendly

Move quickly with a command-line interface, live reloads, and caching. No SDKs, cloud dependencies, or logins.

Get Started

Deploy your way

Get started in minutes with our CLI tool, or choose our managed cloud or on-premises enterprise solutions for advanced features and support.

Get Started
promptfoo deployment options

Secure your AI applications today