llm-guard

Tech stack & purpose

LLM Guard is a security toolkit for Large Language Model interactions developed by Protect AI, designed to sanitize inputs, detect harmful language, prevent data leakage, and defend against prompt injection attacks. The project is built in Python and provides both input and output scanners covering concerns like toxicity, prompt injection, code detection, secrets exposure, and bias. LLM Guard is installable via pip and can be deployed as an API or integrated directly with services like ChatGPT, though the project is no longer under active development as it has been archived.

Languages

Contributors

• asofter298 commits
• Oleksandr Yaremchuk108 commits
• 
  dependabot[bot]76 commits · 0 followers
• Jennifer Cwagenberg8 commits
• Abdelrahman Kaseb4 commits
• Friedrich Große4 commits
• Krish Dholakia2 commits
• Marc Jermaine Pontiveros1 commits
• Martin Lellep1 commits
• Daniel Begimher1 commits
• Alexander1 commits
• Yuki Oshima1 commits
• David Tran1 commits
• Oscaner Miao1 commits
• Cybersizemore1 commits
• Badar Ahmed1 commits
• Edwin Maldonado1 commits
• chahakgoswami1 commits
• Shivansh Anand Srivastava1 commits
• viky1 commits
• Emilio1 commits
• Daniel Iliaev1 commits
• Neenad Ingole1 commits
• radjayAI1 commits
• Andrea Ponti1 commits