TLDR
- Strands Agents is a free, open-source toolkit from AWS that helps developers build AI agents using just a model, tools and a prompt.
- It relies on powerful AI models like Claude or GPT to plan steps, use tools and complete tasks without needing detailed instructions.
- Strands works locally or in the cloud, comes with ready-made tools and is supported by companies like Meta and Anthropic.
AWS has launched Strands Agents, a software development kit designed to make it much easier for developers to build AI agents that can handle complex tasks on their own. It’s open-source and has the potential to simplify how AI is used across everything from customer service bots to software development tools.
Strands taps today’s powerful large language models (LLMs) that can now do more than just answer questions or write text. They can plan, make decisions, and use online tools to get things done – almost like digital assistants that think and act on their own.
In the past, building these kinds of AI agents was difficult and time-consuming. Developers had to carefully tell the AI what steps to take, when to take them, and how to use different tools to finish a task. It could take months to get an agent ready for use in the real world.
“Strands simplifies agent development by embracing the capabilities of state-of-the-art models to plan, chain thoughts, call tools, and reflect,” said Clare Liguori, senior principal software engineer, in a blog post. “With Strands, developers can simply define a prompt and a list of tools in code to build an agent, then test it locally and deploy it to the cloud.”
Strands lets the AI agent figure out the steps itself. All developers need to do is the following:
- Pick an AI model like Claude or GPT.
- Choose some tools the agent can use (like a calendar or file search).
- Write a prompt for the task.
From there, the agent handles the rest – thinking, planning, and acting based on the tools it has and the task it’s given.
“We found that relying on the latest models’ capabilities to drive agents significantly reduced our time to market and improved the end user experience, compared to building agents with complex orchestration logic,” Liguori said.
“Where it used to take months for Q Developer teams to go from prototype to production with a new agent, we’re now able to ship new agents in days and weeks with Strands,” she added.
The name Strands refers to the way the platform connects the AI model and the tools it uses – like two strands of DNA working together. This connection allows the AI to keep looping through steps: thinking, checking its progress, using tools, and adjusting its actions until it finishes the task.
For example, if you asked an AI agent built with Strands to help name your new business, it could check if website domains and social media names are available and then suggest options that fit. And it could do that all by itself, using tools connected to the internet.
Strands even supports more advanced cases, like multiple AI agents working together to handle a big project, or agents that search huge tool libraries to find just the right one to use at the right time.
Strands isn’t just for tech giants like Amazon. Anyone – from small startups to independent developers – can use it. The tool supports a wide range of AI models and can run on your laptop or in the cloud. It even comes with ready-made tools to help developers get started quickly.
Behind the scenes, AWS has already used Strands to build AI features for its own products, including Amazon Q Developer and AWS Glue. And now it’s inviting the wider developer community to build their own AI-powered apps using the same system.
Strands represents a big shift in how AI systems are built. Rather than trying to program every step, developers can now give the AI the tools and let it figure things out.
For everyday users, this could mean smarter apps, faster customer support, better virtual assistants, and AI that feels more helpful and human. As more companies adopt this approach, there will likely be a new generation of AI systems that can handle everything from planning trips to managing workflows – with less effort and more flexibility than ever before.
Strands Agents is available now on GitHub for anyone to try. For those interested in building their own AI agents, the project also includes example code, deployment tools, and community support from the likes of Accenture, Anthropic, Langfuse, Meta, PwC, mem0.ai, Ragas.io, and Tavily.
Be First to Comment