Your agent looks up orders and remembers conversations but invents policies when customers ask about returns or shipping. In this episode we build a searchable knowledge base with embeddings and pgvector so the answers come from real documentation.

Laravel News

A tutorial on building a RAG (Retrieval-Augmented Generation) knowledge base in Laravel 13 using text embeddings and PostgreSQL's pgvector extension. It covers creating a KnowledgeArticle model with a 1,536-dimension vector column, generating embeddings via the Laravel AI SDK's Embeddings facade, seeding FAQ articles, and using Laravel 13's whereVectorSimilarTo method for semantic search. The knowledge base is then wired into a support agent so policy questions return answers from real documentation rather than hallucinated responses.

Ship AI with Laravel: RAG with Embeddings and pgvector in Laravel 13