Learn how to use PyRITβs components to build red teaming workflows.
Load, create, and manage seed datasets for red teaming campaigns.
Run single-turn and multi-turn attacks β Crescendo, TAP, Skeleton Key, and more.
Connect to OpenAI, Azure, Anthropic, HuggingFace, HTTP endpoints, and custom targets.
Transform prompts with text, audio, image, and video converters.
Evaluate AI responses with true/false, Likert, classification, and custom scorers.
Track conversations, scores, and attack results with SQLite or Azure SQL.
Initialize PyRIT, configure defaults, and manage resiliency settings.
Run standardized evaluation scenarios at scale across harm categories.
Register and discover targets, scorers, and converters via class and instance registries.
Use CoPyRIT β the graphical interface for human-led red teaming.
Use pyrit_scan and pyrit_shell for automated and interactive assessments.