RAGFlow
Open-source RAG engine built on deep document understanding — handles tables, figures, and complex layouts in PDFs that chunk-based RAG misses. Includes a visual DAG pipeline editor for building extraction and retrieval workflows. One of the fastest-growing RAG projects of 2025 with 40K+ GitHub stars.
Haystack
deepset's open-source framework for building production NLP and LLM pipelines. Strong focus on RAG, search, and document AI use cases.