Strands is a new Python SDK from Amazon that aims to shorten the time required for developing agentic AI applications.
If MCP wasn't enough, now there's Agents everywhere! Agentic AI promises to go beyond simple MCP cases and enable task completions with complex coordinated workflows. As such, SDKs have started to appear to help build such agents.
AWS is no stranger to agentic tooling. It already had Bedrock Agents and AgentCore. So what makes Strands different? Strands scope as officially stated is to use the toolikt for:
Building production-ready, multi-agent AI systems in a few lines of code
And few lines of code it is, since the simplest agent of all can be made simply by:
First, installing the Strands Agents SDK:
pip install strands-agents
Second, create your first agent as a Python file, i.e. agent.py.
from strands import Agent
# Create an agent with default settings agent = Agent()
# Ask the agent a question agent("Tell me about agentic AI")
Third and finally, run the agent with:
python -u agent.py
After configuring the credentials you can now deploy this on AWS Lambda, Fargate, or EC2.
To answer the question of what's different to the other agentic toolikts of AWS, Strands is fully self-hostable, is open-source, integrates directly with the MCP protocol and allows for building fully customized agents. A high-level overview summarizing its key features is:
Lightweight and stays out of your way: A simple agent loop that just works and is fully customizable.
Production ready: Full observability, tracing, and deployment options for running agents at scale.
Model, provider, and deployment agnostic: Strands supports many different models from many different providers.
Powerful built-in tools: Get started quickly with tools for a broad set of capabilities.
Multi-agent and autonomous agents: Apply advanced techniques to your AI systems like agent teams and agents that improve themselves over time.
Conversational, non-conversational, streaming, and non-streaming: Supports all types of agents for various workloads.
Safety and security as a priority: Run agents responsibly while protecting data.
Strands main selling points can be considered those three:
Model: Strands offers flexible model support. You can use any model in Amazon Bedrock, Anthropic’s Claude model, a model from the Llama model family via Llama API, Ollama for local development and so on.
Tools: You can choose from thousands of published Model Context Protocol (MCP) servers to use as tools for your agent. Strands also provides 20+ pre-built example tools, including tools for manipulating files, making API requests, and interacting with AWS APIs. You can easily use any Python function as a tool, by simply using the Strands @tool decorator.
Prompt: You provide a natural language prompt that defines the task for your agent, such as answering a question from an end user. You can also provide a system prompt that provides general instructions and desired behavior for the agent.
On the project's official repository you'll find a nice collection of agentic use cases and code examples:
Agents Workflows - Example of a sequential agent workflow pattern
CLI Reference Agent - Example of Command-line reference agent implementation
File Operations - Example of agent with file manipulation capabilities
MCP Calculator - Example of agent with Model Context Protocol capabilities
Meta Tooling - Example of Agent with Meta tooling capabilities
Multi-Agent Example - Example of a multi-agent system
Weather Forecaster - Example of a weather forecasting agent with http_request capabilities
Browsing through them, you'll find that the code is very compact, which comes as no surprise since the SDK encapsulates a lot of functionality. That, of course, means that Strands can really shorten the time required for developing your agentic AI applications. Recommended.
A German/US company is working on a hybrid made up of live cockroaches and AI control to produce bio-robotic systems, aka fully controllable living insects, that could be used for "mission-critical op [ ... ]
