Your RAG pipeline is missing two-thirds of the picture

Most RAG pipelines do one thing well: find text chunks that are semantically similar to a query. But a real customer question like "My API calls are failing and I need to upgrade my plan" isn't answered by similarity alone. You need: Semantic similarity to understand the intent (vector search) Keyword precision to catch exact terms like "API" and "upgrade" (text search) Metadata filtering to surface only relevant, high-quality articles (SQL conditions) Relationship awareness to follow the thread from "API rate limits" to "plan upgrade" to "billing FAQ" (graph traversal) Gluing Pinecone + Elasticsearch + Neo4j together to get this is an infrastructure nightmare. VelesQL does it in one query string, inside a single embedded engine. In this tutorial, we'll build a customer support AI agent that uses all three query types against VelesDB v1.9.2. What is VelesQL? VelesQL is a SQL-like language with two domain-specific extensions: NEAR - vector similarity search (HNSW) MATCH - full-text sear