Session

“Search” and “memory” are major factors that determine the accuracy of generative AI apps and AI agents. In this session, we introduce TiDB X, which lets you handle everything from RAG to agent memory in a single database, and explain its strengths as a hybrid RAG foundation that supports both vector search and full-text search (FTS) using standard SQL. In the latter half, we present implementation patterns for AI agent “memory”—how to accumulate and reference conversation history, user information, and long-term memory—with concrete examples and a demo.