What problem does it solve? Building RAG into an Agno agent requires choosing a vector database, wiring embeddings, and keeping infrastructure concerns out of domain code. This Skill guides you through adding searchable knowledge to an agent with the vector backend selected entirely by environment variables, so the same code runs locally and in production. ## Core Features & Use Cases - Pluggable vector backends: Switch between LanceDB (local, zero-infra), PgVector (Postgres-based), and S3Vectors (AWS-native) via the RAG_VECTOR_DB environment variable without touching domain or application code. - Knowledge construction: Create an Agno Knowledge object with OpenAI embeddings, hybrid search, and document loading from URLs, PDFs, and files. - Local observability and evaluation: Inspect retrieval in AgentOS, Langfuse, and SigNoz, and measure effectiveness with golden sets, recall@k, Ragas, and DeepEval. - Use Case: You are building a support agent that must answer from a product manual. Develop locally with LanceDB, then deploy to production on PgVector by changing only environment variables, and validate retrieval quality with a golden question set. ## Quick Start Ask the AI to add RAG to your Agno agent using LanceDB locally with an environment-variable switch to PgVector for production.
