RAG-Anything: All-in-One Multimodal Document Processing and Retrieval Framework
Summary
RAG-Anything is an all-in-one multimodal RAG framework built on LightRAG that ingests, parses, analyzes, and retrieves across text, images, tables, and equations within a single pipeline. It emphasizes unified multimodal processing, a knowledge-graph-based retrieval layer, and extensible parsers (MinerU, DocLing, PaddleOCR) with options for end-to-end LLM-driven querying and multimodal content injection.