How Anthropic wants you to prompt Claude

What it does This repository houses Anthropic’s official curriculum for developers working with Claude. It contains five sequential courses—built as Jupyter Notebook files—that start with Anthropic API fundamentals like keys, parameters, and streaming, then advance through prompt engineering, real-world prompting, production evaluations, and tool use. The examples often favor Claude 3 Haiku so students can follow along without racking up significant API charges.

The interesting bit Most AI vendors publish scattered blog posts; Anthropic turned its guidance into a structured syllabus, with select courses also mirrored on AWS Workshops and Google Vertex. It is unusually candid about its own economics, openly admitting it favors its cheapest model to keep your costs down.

Key highlights

• Five progressive courses covering the full Claude integration lifecycle
• Jupyter Notebook format for interactive, executable examples
• Explicitly optimized for low API spend (often uses Claude 3 Haiku)
• Select courses available via AWS Workshops and Google Vertex AI
• 22,055 stars, indicating broad developer interest

Caveats

• The top-level README is strictly a table of contents; the actual notebooks and depth live one directory down
• Content is tightly bound to Anthropic’s SDK and model family; portability to other LLMs is not the goal

Verdict Worth bookmarking if you are building on Claude and want vendor-tuned best practices rather than generic prompting advice. Skip it if you are looking for model-agnostic theory or a drop-in library.