Add vector search, RAG, and AI agents to your Oracle applications without leaving the database. Most AI tutorials want you to learn Python, run a separate vector store, and move your data out of Oracle. This book skips all that. Everything is built in SQL, PL/SQL, and Oracle APEX 26.1 on Oracle AI Database 26ai, right next to the data you already have. All 237 examples were run in a real lab, and the book prints the output each one returned, including the answers from the language models. Nothing is made up. What you will learn: - The VECTOR type, distance metrics, and similarity search in SQL - Embeddings inside the database with an ONNX model, and from a provider like Gemini - HNSW and IVF vector indexes, semantic search, and hybrid search - Turning PDF, Word, and HTML documents into chunks and embeddings - Calling language models from PL/SQL with DBMS_VECTOR_CHAIN - RAG with citations, and natural language to SQL with guard rails - The APEX_AI package, AI assistants, and AI agents that confirm before they act - Running models locally with Ollama - Security, privacy, answer quality, cost, speed, and deployment You also build three complete projects in Oracle APEX 26.1: a knowledge assistant, an AI help desk, and plain language reporting. Every example uses one sample application, Atlas Support, and the whole lab is free to set up. 29 chapters, six appendices, 47 figures, and all the code free on GitHub. Written for Oracle developers who know SQL and PL/SQL. No machine learning or Python background needed. About the author: Vinish Kapoor is an Oracle ACE Pro with more than twenty years of experience in Oracle SQL, PL/SQL, and APEX. He writes at vinish.dev.
