Strands Agents

from strands import Agent
 
agent = Agent()
agent("What is the capital of France?")

from strands import Agent, tool
 
@tool
def weather(city: str) -> str:
    """Get current weather for a city.
 
    Args:
        city: Name of the city to check weather for.
 
    Returns:
        Current weather description.
    """
    # In production, call a real API
    return f"Sunny, 22°C in {city}"
 
@tool
def convert_temperature(celsius: float) -> float:
    """Convert Celsius to Fahrenheit.
 
    Args:
        celsius: Temperature in Celsius.
 
    Returns:
        Temperature in Fahrenheit.
    """
    return celsius * 9/5 + 32
 
agent = Agent(tools=[weather, convert_temperature])
agent("What's the weather in Madrid? Give me the temperature in Fahrenheit too.")

from strands import Agent
from strands.models import BedrockModel
 
# Amazon Bedrock (default)
agent = Agent(model="anthropic.claude-sonnet-4-20250514-v1:0")
 
# With explicit configuration
bedrock = BedrockModel(
    model_id="anthropic.claude-sonnet-4-20250514-v1:0",
    region_name="us-west-2",
    temperature=0.3,
)
agent = Agent(model=bedrock)

# OpenAI
from strands.models import OpenAIModel
agent = Agent(model=OpenAIModel(model_id="gpt-4o"))
 
# Anthropic direct
from strands.models import AnthropicModel
agent = Agent(model=AnthropicModel(model_id="claude-sonnet-4-20250514"))
 
# Ollama (local)
from strands.models import OllamaModel
agent = Agent(model=OllamaModel(model_id="llama3.1"))

from strands import Agent
from strands.tools.mcp import MCPClient
 
# Connect to an MCP server
mcp = MCPClient(command="uvx", args=["my-mcp-server"])
 
with mcp:
    agent = Agent(tools=mcp.list_tools())
    agent("Use the MCP tools to complete this task")

from strands import Agent
 
researcher = Agent(
    system_prompt="You are a research specialist.",
    tools=[web_search]
)
 
writer = Agent(
    system_prompt="You are a technical writer.",
    tools=[researcher.as_tool(
        name="research",
        description="Research a topic thoroughly"
    )]
)
 
writer("Write an article about quantum computing")

from strands import Agent
from strands.multiagent import Swarm
 
triage = Agent(system_prompt="Route to the right specialist.")
billing = Agent(system_prompt="Handle billing questions.", tools=[billing_api])
technical = Agent(system_prompt="Handle technical issues.", tools=[diagnostics])
 
swarm = Swarm(
    agents={"triage": triage, "billing": billing, "technical": technical},
    entry_point="triage"
)
 
swarm("I can't log in and I was charged twice")

from strands.multiagent import Graph
 
graph = Graph()
graph.add_node("classify", classify_agent)
graph.add_node("respond", respond_agent)
graph.add_node("escalate", escalate_agent)
 
graph.add_edge("classify", "respond", condition=lambda r: r.priority == "low")
graph.add_edge("classify", "escalate", condition=lambda r: r.priority == "high")
 
graph.run("Customer complaint about delivery")

result = agent("What is the square root of 144?")
 
# Usage metrics
print(result.metrics.get_summary())
# {
#   "total_cycles": 2,
#   "total_duration": 1.88,
#   "accumulated_usage": {"inputTokens": 3921, "outputTokens": 83},
#   "tool_usage": {"calculator": {"call_count": 1, "success_rate": 1.0}}
# }

import asyncio
from strands import Agent
 
agent = Agent(callback_handler=None)
 
async def stream():
    async for event in agent.stream_async("Explain quantum computing"):
        if "data" in event:
            print(event["data"], end="", flush=True)
 
asyncio.run(stream())

from pydantic import BaseModel
from strands import Agent
 
class MovieReview(BaseModel):
    title: str
    rating: float
    summary: str
 
agent = Agent()
result = agent.structured_output(
    "Review the movie Inception",
    output_model=MovieReview
)
print(result.title)   # "Inception"
print(result.rating)  # 9.2