The harness behind Claude Code, packaged so you can build on it. It ships real tools: read and write files, run bash, search a codebase, fetch the web. You give it a prompt and options, and it drives the loop on infrastructure you host.
Reach for it when
- You are building an agent that works on a filesystem or a repo
- You want strong built in tools rather than writing your own
- You want Claude Code behaviour inside your own product
Look elsewhere when
- Your agent never touches files or a shell
- You would rather host nothing at all
What the lessons will cover
- 01Install and authenticate
- 02Your first query
- 03Built in tools
- 04Permissions and approval
- 05Custom tools with MCP
- 06Subagents
- 07Hooks
- 08Shipping one
Others in vendor sdk
Google ADK
OpenAI Agents SDKOpenAI's own agent loop, with handoffs and guardrails.
Semantic KernelMicrosoft's SDK for wiring models into real applications.
Vercel AI SDKThe TypeScript way to build streaming AI apps.Common questions
Is Claude Agent SDK free to use?
Claude Agent SDK is open source and free to run yourself. You still pay whichever model provider you point it at, and this tutorial is free with no signup.
Do I need to know Python to use Claude Agent SDK?
Basic Python is enough. Functions, dictionaries and imports cover most of what Claude Agent SDK asks of you.
When should I not use Claude Agent SDK?
Your agent never touches files or a shell. You would rather host nothing at all.
Lessons for Claude Agent SDK are being written. Meanwhile the LangGraph tutorial covers the same ground: state, tools, loops, memory and human approval. Most of it carries straight over.
Start the LangGraph tutorial →