Add a Vector Database to Your FastAPI AI App

In the previous workshop, we focused on turning an AI agent or RAG-style prototype into a structured FastAPI backend.

We covered the backend setup, API structure, configuration, and the foundation for serving an AI application through clean endpoints. But we didn’t have enough time to properly cover one important part: adding a vector database.

So this session is Part 2.

In this hands-on workshop, we’ll continue from the previous project and focus specifically on the retrieval layer. You’ll learn how to connect a vector database, ingest and index documents, retrieve relevant context, and use that context to generate grounded responses through your FastAPI app.

What you’ll learn

• How to add a vector database to an existing FastAPI AI application
• How to prepare documents for indexing
• How to generate and store embeddings
• How to query the vector database for relevant context
• How to connect retrieval results to an LLM response
• How to structure the ingestion and retrieval logic so it can be extended later
• What to consider when using vector search in a production-style setup

This session is for people who joined the first workshop or already have a basic FastAPI AI backend and want to add vector search to it.

You don’t need a production-ready app, but you should be comfortable with basic Python and API concepts.