neo4j-labs/llm-graph-builder

This project is an advanced application designed to automate the construction of knowledge graphs from a wide array of unstructured data sources, including documents, web pages, and videos. By integrating Large Language Models (LLMs) and the LangChain framework, it efficiently extracts entities, relationships, and properties, storing them as structured graphs in Neo4j. Key functionalities include customizable schema application, visualization through Neo4j Bloom, and a conversational interface for querying the generated graphs. It also provides robust token usage tracking and supports a variety of LLM and embedding models, making complex data accessible and actionable for various analytical and AI applications.