PyRIT
Python Risk Identification Tool for red-teaming generative AI.
Microsoft’s open-source framework for automating red-team attacks — jailbreaks, prompt injection, harmful content — against models and AI apps, with scorers and multi-turn attack strategies.
- Vendor
- Microsoft
- Category
- AI security & guardrails
- Pricing
- Open source
- Open source
- Yes
- License
- MIT
- Platforms
- Python
- Launched
- 2024
- Website
- microsoft.github.io/PyRIT
Features
- Attack orchestrators
- Multi-turn strategies
- Scorers
Best for
- Enterprise
More ai security & guardrails
- NeMo Guardrails — NVIDIA. Open-source toolkit for programmable LLM guardrails.
- Guardrails AI — Guardrails AI. Validators that check and correct LLM inputs and outputs.
- Amazon Bedrock Guardrails — Amazon Web Services. Managed safety policies for generative AI on AWS.
- Lakera Guard — Lakera (Check Point). Real-time prompt-injection and data-leak protection API.
- garak — NVIDIA. LLM vulnerability scanner for red-teaming.
- Model Armor — Google Cloud. Screens prompts and responses for injection, data leaks and harmful content.
- LlamaFirewall — Meta. Open-source guardrail system for AI agents.
- LLM Guard — Protect AI (Palo Alto Networks). Archived: open-source scanners for LLM prompts and outputs.