Back to Details

aws-strands

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Strands Agents SDK

Strands Agents SDK

Build model-agnostic AI agents with the Strands framework.
使用Strands框架构建模型无关的AI Agent。

Installation

安装

bash
pip install strands-agents strands-agents-tools
bash
pip install strands-agents strands-agents-tools

Or with npm

Or with npm

npm install @strands-agents/sdk
undefined
npm install @strands-agents/sdk
undefined

Quick Start

快速开始

python
from strands import Agent
from strands.tools import tool

@tool
def get_weather(city: str) -> str:
    """Get current weather for a city."""
    # Implementation
    return f"Weather in {city}: 72°F, Sunny"

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[get_weather]
)

response = agent("What's the weather in Seattle?")
print(response)
python
from strands import Agent
from strands.tools import tool

@tool
def get_weather(city: str) -> str:
    """Get current weather for a city."""
    # Implementation
    return f"Weather in {city}: 72°F, Sunny"

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[get_weather]
)

response = agent("What's the weather in Seattle?")
print(response)

TypeScript/JavaScript

TypeScript/JavaScript

typescript
import { Agent, tool } from '@strands-agents/sdk';

const getWeather = tool({
  name: 'get_weather',
  description: 'Get current weather for a city',
  parameters: {
    city: { type: 'string', description: 'City name' }
  },
  handler: async ({ city }) => {
    return `Weather in ${city}: 72°F, Sunny`;
  }
});

const agent = new Agent({
  model: 'anthropic.claude-3-sonnet',
  tools: [getWeather]
});

const response = await agent.run('What\'s the weather in Seattle?');
typescript
import { Agent, tool } from '@strands-agents/sdk';

const getWeather = tool({
  name: 'get_weather',
  description: 'Get current weather for a city',
  parameters: {
    city: { type: 'string', description: 'City name' }
  },
  handler: async ({ city }) => {
    return `Weather in ${city}: 72°F, Sunny`;
  }
});

const agent = new Agent({
  model: 'anthropic.claude-3-sonnet',
  tools: [getWeather]
});

const response = await agent.run('What\'s the weather in Seattle?');

Model Agnostic

模型无关性

Strands works with any LLM:
python
from strands import Agent
Strands可与任意LLM兼容：
python
from strands import Agent

Anthropic (default)

Anthropic (default)

agent = Agent(model="anthropic.claude-3-sonnet")
agent = Agent(model="anthropic.claude-3-sonnet")

OpenAI

OpenAI

agent = Agent(model="openai.gpt-4o")
agent = Agent(model="openai.gpt-4o")

Amazon Bedrock

Amazon Bedrock

agent = Agent(model="amazon.titan-text-premier")
agent = Agent(model="amazon.titan-text-premier")

Custom endpoint

Custom endpoint

agent = Agent( model="custom", endpoint="https://your-model-endpoint.com", api_key="..." )
undefined
agent = Agent( model="custom", endpoint="https://your-model-endpoint.com", api_key="..." )
undefined

Tool Definition Patterns

工具定义模式

Decorator Style

装饰器风格

python
from strands.tools import tool

@tool
def search_database(query: str, limit: int = 10) -> list[dict]:
    """Search the product database.
    
    Args:
        query: Search query string
        limit: Maximum results to return
    """
    # Implementation
    return results
python
from strands.tools import tool

@tool
def search_database(query: str, limit: int = 10) -> list[dict]:
    """Search the product database.
    
    Args:
        query: Search query string
        limit: Maximum results to return
    """
    # Implementation
    return results

Class Style

类风格

python
from strands.tools import Tool

class DatabaseSearchTool(Tool):
    name = "search_database"
    description = "Search the product database"
    
    def parameters(self):
        return {
            "query": {"type": "string", "description": "Search query"},
            "limit": {"type": "integer", "default": 10}
        }
    
    def run(self, query: str, limit: int = 10):
        return self.db.search(query, limit)
python
from strands.tools import Tool

class DatabaseSearchTool(Tool):
    name = "search_database"
    description = "Search the product database"
    
    def parameters(self):
        return {
            "query": {"type": "string", "description": "Search query"},
            "limit": {"type": "integer", "default": 10}
        }
    
    def run(self, query: str, limit: int = 10):
        return self.db.search(query, limit)

ReAct Pattern

ReAct模式

Built-in ReAct (Reasoning + Acting) support:
python
from strands import Agent, ReActStrategy

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[search_tool, calculate_tool],
    strategy=ReActStrategy(
        max_iterations=10,
        verbose=True
    )
)
内置ReAct（推理+行动）支持：
python
from strands import Agent, ReActStrategy

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[search_tool, calculate_tool],
    strategy=ReActStrategy(
        max_iterations=10,
        verbose=True
    )
)

Agent will reason through complex multi-step tasks

Agent will reason through complex multi-step tasks

response = agent(""" Find the top 3 products in our database, calculate their average price, and recommend if we should adjust pricing. """)
undefined
response = agent(""" Find the top 3 products in our database, calculate their average price, and recommend if we should adjust pricing. """)
undefined

Multi-Agent Systems

多Agent系统

python
from strands import Agent, MultiAgentOrchestrator
python
from strands import Agent, MultiAgentOrchestrator

Specialist agents

Specialist agents

researcher = Agent( name="researcher", model="anthropic.claude-3-sonnet", tools=[web_search, document_reader], system_prompt="You are a research specialist." )
analyst = Agent( name="analyst", model="anthropic.claude-3-sonnet", tools=[data_analyzer, chart_generator], system_prompt="You are a data analyst." )
writer = Agent( name="writer", model="anthropic.claude-3-sonnet", tools=[document_writer], system_prompt="You are a technical writer." )
researcher = Agent( name="researcher", model="anthropic.claude-3-sonnet", tools=[web_search, document_reader], system_prompt="You are a research specialist." )
analyst = Agent( name="analyst", model="anthropic.claude-3-sonnet", tools=[data_analyzer, chart_generator], system_prompt="You are a data analyst." )
writer = Agent( name="writer", model="anthropic.claude-3-sonnet", tools=[document_writer], system_prompt="You are a technical writer." )

Orchestrator

Orchestrator

orchestrator = MultiAgentOrchestrator( agents=[researcher, analyst, writer], routing="supervisor" # or "round_robin", "intent" )
response = orchestrator.run( "Research AI trends, analyze the data, and write a report" )
undefined
orchestrator = MultiAgentOrchestrator( agents=[researcher, analyst, writer], routing="supervisor" # or "round_robin", "intent" )
response = orchestrator.run( "Research AI trends, analyze the data, and write a report" )
undefined

Streaming Responses

流式响应

python
from strands import Agent

agent = Agent(model="anthropic.claude-3-sonnet")
python
from strands import Agent

agent = Agent(model="anthropic.claude-3-sonnet")

Stream response

Stream response

for chunk in agent.stream("Explain quantum computing"): print(chunk, end="", flush=True)
undefined
for chunk in agent.stream("Explain quantum computing"): print(chunk, end="", flush=True)
undefined

Memory Management

内存管理

python
from strands import Agent
from strands.memory import ConversationMemory, SemanticMemory

agent = Agent(
    model="anthropic.claude-3-sonnet",
    memory=[
        ConversationMemory(max_turns=10),
        SemanticMemory(embedding_model="text-embedding-3-small")
    ]
)
python
from strands import Agent
from strands.memory import ConversationMemory, SemanticMemory

agent = Agent(
    model="anthropic.claude-3-sonnet",
    memory=[
        ConversationMemory(max_turns=10),
        SemanticMemory(embedding_model="text-embedding-3-small")
    ]
)

Memory persists across calls

Memory persists across calls

agent("My name is Alice") agent("What's my name?") # Remembers: "Your name is Alice"
undefined
agent("My name is Alice") agent("What's my name?") # Remembers: "Your name is Alice"
undefined

AgentCore Integration

AgentCore集成

Use Strands with AWS Bedrock AgentCore:
python
from strands import Agent
from strands.tools import tool
import boto3

agentcore_client = boto3.client('bedrock-agentcore')

@tool
def query_cloudwatch(metric_name: str, namespace: str) -> dict:
    """Query CloudWatch metrics via AgentCore Gateway."""
    return agentcore_client.invoke_tool(
        tool_name="cloudwatch_query",
        parameters={"metric": metric_name, "namespace": namespace}
    )

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[query_cloudwatch]
)
将Strands与AWS Bedrock AgentCore配合使用：
python
from strands import Agent
from strands.tools import tool
import boto3

agentcore_client = boto3.client('bedrock-agentcore')

@tool
def query_cloudwatch(metric_name: str, namespace: str) -> dict:
    """Query CloudWatch metrics via AgentCore Gateway."""
    return agentcore_client.invoke_tool(
        tool_name="cloudwatch_query",
        parameters={"metric": metric_name, "namespace": namespace}
    )

agent = Agent(
    model="anthropic.claude-3-sonnet",
    tools=[query_cloudwatch]
)

Official Use Cases

官方用例

Strands is featured in AWS AgentCore samples:
A2A Multi-Agent Incident Response: Uses Strands for monitoring agent
bash
cd amazon-bedrock-agentcore-samples/02-use-cases/A2A-multi-agent-incident-response
Strands已被纳入AWS AgentCore示例：
A2A多Agent事件响应：使用Strands作为监控Agent
bash
cd amazon-bedrock-agentcore-samples/02-use-cases/A2A-multi-agent-incident-response

Monitoring agent uses Strands SDK for CloudWatch, logs, metrics

Monitoring agent uses Strands SDK for CloudWatch, logs, metrics

undefined
undefined

Resources

资源