RAGFlow v0.24.0, released in February 2026, introduced significant enhancements across memory management, knowledge governance, agent capabilities, and document processing. Memory management is a major addition with new HTTP and Python APIs for memory extraction, plus console logging of memory extraction activities, enabling agents to maintain context across conversations. Knowledge base governance improvements include batch metadata management and a terminology update—"Table of Contents" is now called “PageIndex” for clarity. The agent conversation interface received a major upgrade with a new Chat-like interface that retains session and dialogue history, making multi-turn agent interactions feel natural and maintaining context across user interactions. The multi-Sandbox mechanism now supports both local gVisor and Alibaba Cloud sandboxes, providing flexibility in execution environments while maintaining security through isolation—compatibility with mainstream Sandbox APIs enables deployment across different infrastructure. LLM enhancements introduced a new “Thinking” mode (aligned with models like OpenAI’s o1), improving reasoning and research capabilities while removing the previous “Reasoning” configuration option. Retrieval strategies were optimized for deep-research scenarios, enhancing recall accuracy for complex, multi-step questions. Administration features expanded to support multiple Admin accounts (useful for large teams) and added model connection testing for newly configured models to validate connectivity before deployment. Alternative database support grew with OceanBase as an option alongside MySQL for system data storage. Finally, the document engine was upgraded to Infinity v0.6.1, improving parsing speed and accuracy for complex documents. Together, these changes position RAGFlow as a more enterprise-ready, agentic-first platform with stronger conversational abilities and operational governance For production deployments, Milvus provides a dedicated open-source vector database optimized for RAG pipelines, while Zilliz Cloud offers a managed alternative with enterprise-grade performance and reliability…
Related Resources: RAG Pipeline with Milvus | Improving Chunking for RAG