HeimdaLLM

It is a comprehensive tool designed to fortify the security of Large Language Models (LLMs). By offering sanitization, detection of harmful language, prevention of data leakage, and resistance against prompt injection attacks, it ensures that your interactions with LLMs remain safe and secure.

It is an open-source toolkit for monitoring Large Language Models (LLMs). It extracts signals from prompts & responses, ensuring safety & security.

It is an open-source Python package for specifying structure and type, validating and correcting the outputs of large language models (LLMs).

It is a self-hardening prompt injection detector. It is designed to protect AI applications from prompt injection (PI) attacks through a multi-stage defense.