Tool overview
Promptfoo is listed under AI Governance Safety & Evaluation AI tools.
What is Promptfoo?
Promptfoo tests prompts, models, RAG systems, and agents through declarative evaluations, assertions, red teaming, vulnerability scanning, CI/CD workflows, and provider integrations. The MIT-licensed Community edition runs locally, while Enterprise adds managed cloud, monitoring, APIs, collaboration, and on-premises deployment.
Best for
AI developers and security teams testing prompts, agents, and RAG systems
Who is it for?
Decision note
Best for developers and security teams that need repeatable, code-friendly evaluation and adversarial testing across many models and application architectures.
Key features
Declarative prompt, RAG, and agent evaluations
Automated red teaming and vulnerability scanning
Broad provider and CI/CD integrations
Local open-source, managed cloud, and on-premises editions
Use cases
Regression-test prompts and agents
Red-team AI applications before release
Compare models and provider behavior
Automate security evaluations in CI/CD
Pros
- Community edition is free forever
- Open-source code uses the MIT License
- Supports local and self-hosted testing
Limitations
Generated tests and scan findings can miss context-specific risks or produce false positives. Teams must review results, test realistic targets, secure provider credentials, and maintain independent security and compliance processes.
Pricing details
Billing options
Pricing note
Community is free forever and includes core evaluation plus up to 10,000 red-team probes per month. Enterprise and Enterprise On-Premise use custom pricing for collaboration, monitoring, API access, support, SLAs, managed cloud, and isolated deployment.
Supported languages
- English
Integrations
OpenAI
Anthropic
Azure OpenAI
Google models
Hugging Face
Ollama
CI/CD systems
Please log in to join the discussion.