Build Apps with LLM APIs
From your first API call to streaming, structured data and tool-using agents in Python
A hands-on course for developers who want to put large language models inside real software. You will write Python against the Claude API from the very first lesson: keep API keys safe, hold multi-turn conversations, measure tokens and cost, cut bills with prompt caching, stream answers into a web endpoint, extract validated structured data, and give the model tools it can call. The final lesson turns one of those tools into a tiny MCP server that any MCP client can use. Every concept comes with small, runnable code and a graded coding exercise. The patterns carry over to other providers, but the examples use the official Anthropic Python SDK.
6 lessons · ~23 min video
First lesson free · then membership
Course content
Module 1
Talking to a model from code
Set up the SDK and keys, send messages with system prompts and history, and understand what every call costs.
Module 2
Building features
Stream answers, extract structured data, and let the model call your tools, then share a tool with any MCP client.