Until now you have talked to models in a chat window. A RAG system talks to them from code: your application builds a prompt, sends it to a model over an API, and reads the answer back. Knowing how those calls work is what turns a good prompt into a working product.
The lessons start with a beginner tutorial for the OpenAI API, covering a basic chatbot, tools, and streaming. Then two episodes from a Claude API crash course show how to set up a project and send your first message with a second provider, so you can see which ideas are shared across APIs. Next comes tool use and function calling, which lets a model ask your code to do something, and a lesson on structured outputs explains how to get reliable JSON back instead of free-form text.
Watch the lessons in order, and try the code yourself as you go. APIs change quickly, so if a call in a video no longer works, check the provider's current documentation.
A step-by-step introduction to calling the OpenAI API from code: a basic chatbot, image analysis, built-in and custom tools, and streaming the response.
Video 14minGetting started with the Claude API: what you need before you write any code and how to set up a project to call the model.
Video 9minSending your first message to the model and reading the response, the basic request and response cycle behind every LLM application.
Video 9minHow tool use and function calling let a model request an action, such as a search or a calculation, that your code carries out before the model writes its final answer.
Video 7minWhy free-form text is hard to use in software, and how structured outputs make a model return JSON that follows a schema your code can rely on.
Video 9min