jev-llm-guard

An open-source tool providing contextual guardrails for large language models (LLMs) based on the OWASP LLM Top 10 2026 taxonomy. It evaluates textual inputs for security risks like prompt injection and misinformation, then returns allow, review, or block decisions with detailed risk assessments. Includes a Node.js local demo, an interactive Turkish and English demo, and a library for integration.