Claude Platform 101
Anthropic
Last updated October 12, 2026
Claude Platform 101 is a free video course from Anthropic that teaches developers to build on the Claude Platform from the ground up — the step from asking Claude questions in a chat window to building it into software you release. It starts at first principles: what a single API request contains, how to read the response, and how to pick a model for the job (Fable, Opus, Sonnet, or Haiku) by running a quick evaluation rather than guessing. From there it builds the agent loop by hand so you can see how the model acts, observes, and decides, then shows how the Tool Runner in Anthropic's software development kit (SDK) collapses that scaffolding once you understand what it is doing for you. The later lessons cover tool use and extended thinking, Anthropic's built-in tools (web search, code execution, web fetch), Skills that package a procedure once and reuse it, Model Context Protocol (MCP) servers that connect Claude to outside tools without hand-writing a schema, the context-management habits that keep a long-running agent inside its window and affordable, and managed agents that Anthropic runs for you in a sandbox. Most demos use the Python SDK, and a few lessons use the TypeScript SDK with Node and npm. Most lessons come with a code demo. The course itself is free, but running its examples uses the Claude API, which you pay for once a new account's small free credit is used up.
What you'll learn
- Sending your first Claude API request and reading the response
- Choosing a model (Fable, Opus, Sonnet, or Haiku) by running a quick evaluation instead of guessing
- Building the agent loop by hand, then collapsing it with the SDK's Tool Runner
- Giving Claude reach with tool use, and turning on extended thinking for harder problems
- Using Anthropic's built-in tools (web search, code execution, web fetch) and packaging procedures as Skills
- Connecting Claude to outside tools with Model Context Protocol (MCP) servers, without hand-writing schemas
- Managing the context window to keep a long-running agent affordable past the early turns
- Handing the loop to Anthropic with managed agents, and consuming the event stream as it runs
Frequently asked questions about Claude Platform 101
Who is Claude Platform 101 for?
Developers who have used Claude in a chat window and want to build it into their own applications — adding AI features to an existing product, or prototyping an agent from scratch — with no prior agent-building experience assumed.
Is Claude Platform 101 free?
Yes — Claude Platform 101 is free to take. Running its examples uses the Claude API, which you pay for once a new account's small free credit is used up.
What are the prerequisites for Claude Platform 101?
Comfort reading and writing code in at least one language, plus basic command-line familiarity. Most demos use the Python SDK (anthropic, installed with pip), and a few lessons, including the first API call, use the TypeScript SDK (@anthropic-ai/sdk) with Node and npm — you don't need to be expert in either language, but you should be able to follow along and run a script. You'll need a Claude Console account and an API key; new accounts get a small amount of free credit for testing the API, and after that you add credit yourself. No prior experience building with large language models is required.
Do you need to code for Claude Platform 101?
Yes — Claude Platform 101 involves hands-on coding.
Why we suggest this course
A fast, hands-on orientation to building real applications on Claude — not prompting in a chat window. If you want the deeper treatment afterwards, Building with the Claude API is the longer Python deep-dive. Two things worth knowing: it teaches Anthropic's specific platform rather than building with AI in general; and to run the examples you'll need a Claude Console account and an API key.