Big TechPythonTypeScript

Claude Agent SDK

by AnthropicUpdated Jun 15, 2025

Anthropic's production-grade agent runtime with deep MCP integration, computer use capabilities, and a developer-first design philosophy. It provides an agentic loop that automatically handles tool calls, supports multi-turn conversations, and includes built-in guardrails for safe agent behavior.

Architecture Overview

The SDK wraps the Anthropic Messages API in an agentic loop that iterates until the model produces a stop response. Each iteration sends the conversation history (including tool results) back to the model, which decides whether to call more tools or return a final answer. MCP servers are integrated as first-class tool providers via a client-server transport layer.

When to Use Claude Agent SDK

Production agent systems with tool use
Computer use automation (browser, desktop)
MCP-native applications and integrations
Multi-turn conversational assistants
Code generation and analysis pipelines

Strengths & Weaknesses

Strengths

Deep MCP integration as a first-class feature
Computer use support for GUI automation
Strong safety features and content filtering
Clean, Pythonic API with excellent TypeScript support
Built-in support for streaming responses

Weaknesses

Primarily optimized for Anthropic Claude models
Newer ecosystem compared to LangChain/LlamaIndex
Community tooling still catching up to more established frameworks

Quick Start

python

import anthropic

client = anthropic.Anthropic()

# Define tools for the agent
tools = [
    {
        "name": "get_weather",
        "description": "Get the current weather for a location.",
        "input_schema": {
            "type": "object",
            "properties": {
                "location": {"type": "string", "description": "City name"}
            },
            "required": ["location"],
        },
    }
]

def run_agent(user_message: str):
    messages = [{"role": "user", "content": user_message}]

    while True:
        response = client.messages.create(
            model="claude-sonnet-4-20250514",
            max_tokens=1024,
            tools=tools,
            messages=messages,
        )

        # If the model stops, return the final text
        if response.stop_reason == "end_turn":
            return response.content[0].text

        # Process tool calls
        for block in response.content:
            if block.type == "tool_use":
                tool_result = execute_tool(block.name, block.input)
                messages.append({"role": "assistant", "content": response.content})
                messages.append({
                    "role": "user",
                    "content": [{
                        "type": "tool_result",
                        "tool_use_id": block.id,
                        "content": tool_result,
                    }],
                })

result = run_agent("What's the weather in San Francisco?")
print(result)

Features at a Glance

Developer	Anthropic
Language	Python, TypeScript
License	MIT
GitHub Stars	10k+
MCP Support	Yes
Multi-Agent	Yes

Notable Users

AnthropicNotionDuckDuckGoSourcegraph

Resources

Official Documentation

https://docs.anthropic.com/en/docs/agents-and-tools/agent-loop

GitHub Repository

https://github.com/anthropics/anthropic-sdk-python

Explore Related Content

Concept

Edit on GitHub