Augmenting RAG Agents with MCP Servers

In the previous chapters, our agents could call tools and hold conversations — but they could only work with the information the user typed in or the tool returned in real time. What if you want an agent that can answer questions about a 100-page PDF? You can't paste the whole document into the prompt — it's too large. Instead, you need a way to find just the relevant parts and feed only those to the model. That technique is called Retrieval-Augmented Generation, or RAG.