Tool overview
PyRIT is listed under AI Governance Safety & Evaluation AI tools.
What is PyRIT?
PyRIT is Microsoft's MIT-licensed Python Risk Identification Tool for generative AI. It provides targets, converters, scorers, attacks, memory, datasets, a scanner CLI, and the CoPyRIT interface for repeatable adversarial testing.
Best for
Security professionals conducting authorized generative AI risk assessments
Who is it for?
Decision note
Rebuilt from the original export under the complete V411 factual-source-verification workflow. Preview only. Apply remains blocked until Failures = 0, Warnings = 0, Unmapped = 0, Missing = 0, all explicit clears are reviewed, image import is disabled or V380 accepts the asset, one representative WordPress edit screen is compared with the export and proposed row, and a post-Apply zero-change Preview is completed.
Key features
Attack, target, converter, and scorer abstractions
Scanner CLI and CoPyRIT interface
Memory, datasets, and result export
Repeatable single-turn and multi-turn assessments
Use cases
Red-team models and agents
Test prompt-injection resistance
Run repeatable attack campaigns
Export findings for review
Pros
- MIT licensed and actively maintained
- Versioned security patches and release notes
- Built by Microsoft security researchers
Limitations
Use only against systems you own or are authorized to test
attacks can generate harmful, private, or destabilizing content. Results are not complete assurance and require expert interpretation, scope control, cleanup, and responsible disclosure.
Pricing details
Billing options
Pricing note
PyRIT is free under MIT. Costs come from compute, storage, target-model APIs, test environments, and security operations.
Supported languages
- English
Integrations
Azure OpenAI
OpenAI-compatible targets
Hugging Face
Custom targets
Please log in to join the discussion.