Apache Doris(https://doris.apache.org/) now supports vector search, AI functions, MCP, and other AI workloads 🚀 Many reached out wanting to learn more about Apache Doris 4.0 and its new AI features, after we showed them in a recent Agentic AI webinar with AWS. Here is a quick summary of the key AI capability in Apache Doris 4.0: 1️⃣ Vector Search Doris 4.0 supports vector indexing with the HNSW type (Hierarchical Navigable Small Worlds). This essentially allows you to do hybrid search in Doris, combining vector search and regular queries in one SQL code. 2️⃣ AI Functions Doris 4.0 introduces built-in AI Functions that let you call LLMs directly from SQL. You can ask LLMs extract keywords, classify sentiment, summarize text, and more, all in SQL queries. Supporting OpenAI, Anthropic, Gemini, DeepSeek, and more LLMs. 3️⃣ Hybrid Search and Analytics Processing (HSAP) Doris also support powerful full-text search, and introduce the BM25 scoring to better match keywords and phrases. Doris 4.0 now unifies full-text search, vector search, and structured analytics in one engine. This means you can mix semantic search, keyword matching, and precise SQL conditions in the same query, without sending data to separate systems. 4️⃣ MCP Server We also built Doris MCP Server, connecting AI system directly with Doris data. The MCP server helps AI agents and LLMs to have real-time data access and analysis on data stored on Doris. 👉 In short, with Apache Doris 4.0, you can do vector search alongside regular analytical queries in the same SQL engine, without managing multiple databases, getting one simple architecture for analytics & search + AI.