//
Search across all documentation pages
11 pages in this section.
Learn to use the Claude Agent SDK with 9 Python examples, covering basic queries, model selection, and handling streamed responses.
Learn the Claude Agent SDK's mental model, including its tool-use loop, built-in tools, subagents, and sessions, to build production AI agents.
Enable and scope built-in tools like file editing, bash, web search, and web fetch in the Claude Agent SDK. Learn to configure permissions for safe agent operation.
Learn to use subagents in the Claude Agent SDK for context isolation and parallelism. Delegate tasks, define subagents, and manage tool scoping and results.
Reference Claude Agent SDK configuration options for Python and TypeScript. Find names, types, and defaults for model, tool access, and session settings.
Connect the Claude Agent SDK to MCP servers. Register stdio servers for local processes or HTTP servers for remote tools.
Learn best practices for deploying Claude Agent SDK agents safely in production, covering tool scoping, checkpoints, subagents, and session persistence.
A single-page roundup of every highlight bullet from the 10 pages in the Claude Agent SDK section, grouped by source page so you can scan all 49 takeaways without opening each article individually.