Section 7: LLM access via API

Under every assistant and agent is an API call. This section takes you underneath the chat window: getting an API key, understanding how ellmer talks to a provider, building a simple chat and a DIY chatbot, then building your own coding agent from scratch — writing the tool-calling loop by hand so nothing is hidden — running experiments with it to compare models and prompts, and finally applying all of it to a literature review example.