Claude Agent SDK: Build Powerful AI Agents with Claude

The Claude Agent SDK is your toolkit for creating sophisticated AI agents that can autonomously interact with your environment. Think of it as giving Claude the ability to read files, execute commands, browse the web, and even modify code – all programmatically. It leverages the same advanced agent loop and context management that powers Claude Code, making it easier than ever to build intelligent agents.

Claude Agent SDK Key Features

Built-in Tooling: The SDK comes pre-equipped with essential tools for common agent tasks. This includes capabilities for reading files, running shell commands, and editing code directly, so your agent can get to work right away without extensive setup.
Programmable in Python & TypeScript: Whether you prefer Python or TypeScript, the Agent SDK offers robust libraries that integrate seamlessly into your existing development workflows.
Autonomous Agent Loop: Design agents that can reason, plan, and execute multi-step tasks. The SDK handles the complex agent loop for you, allowing you to focus on defining your agent's goals and capabilities.
Flexible Tool Integration: Easily configure your agents to use specific tools from a range of options, including web search, code execution, and file manipulation, enabling customized agent behavior.
Environment Variable Authentication: Securely connect your agent to Claude's powerful models by setting your API key as an environment variable. The SDK also supports authentication through major cloud providers like Amazon Bedrock, Google Vertex AI, and Microsoft Azure.