Build a Local AI Agent in Python (Part 2)

A step-by-step tutorial on enhancing a local Python AI agent with dynamic context injection. Using a decorator pattern, context functions are registered to supply fresh information (current date/time, current user) to the LLM on every request rather than just once at startup. The approach is demonstrated with a locally running model (Qwen 3.5 via LM Studio), showing how context functions can be any regular Python function — reading files, querying databases, or calling APIs. The tutorial builds a 90-line interactive chat loop and previews adding tool-use capabilities in a future part.