This is a framework for building evaluation, testing and monitoring LLM application platforms. It provides an intuitive web interface and can record all LLM calls during development and production. It supports functions such as feedback scoring, test case storage, hallucination detection and CI/CD integration. It is suitable for application scenarios such as RAG chat machines, code assistants and complex agent pipelines.