Skip to main content

No prior API experience assumed

Start with one Claude request. End ready to build.

This is a hands-on path for developers who are new to Anthropic and the Claude API. You will run small pieces of real code, change one thing at a time, and learn why each production pattern exists before you need it.

Start hereNode, an API key, and curiosity. No AI background needed.
Learn by changingEvery stage gives you a small working change, not a lecture to memorize.
Finish with judgmentUse the mission to combine the patterns into a system you can explain.
Make your first request

Short on time? The Tutor builds a timed cram plan from these labs and reviews your exercises.

The only mental model you need first

You send messages; Claude sends a response.

A model is the engine that generates the response. The SDK sends your request to Anthropic’s Messages API and gives your program a response object back.

const response = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Explain zippers." }],
});
modelWhich Claude engine should answer.
messagesWhat your application asks Claude to do.
responseContent, stop reason, and token usage to inspect.

Your path

Build confidence one working layer at a time.

Do these in order. Each step assumes only what the previous one taught.

01

Make your first request

Run one small Claude call. Learn what a model, message, response, and token are.

20 min →02

Turn an answer into data

Give Claude a schema so your code receives a classification it can safely use.

35 min →03

Let Claude check facts

Connect tools, watch the model ask for order and policy data, and inspect its trace.

45 min →04

Make the experience feel alive

Stream a response to the browser and handle the ending and error states correctly.

30 min →05

Learn to operate it

Measure cost, caching, model choice, and safety before you call something production-ready.

2–3 hrs →