Amazon Web Services (AWS) unveiled Strands Agents SDK, a new open-source software development tool for creating artificial intelligence (AI) agents.
The Strands Agents SDK utilizes a model-driven approach that maximizes the inference and tool usage capabilities of cutting-edge large language models (LLMs) without requiring AI orchestration.
This innovative tool allows developers to rapidly build and deploy production-grade AI agents with minimal coding effort.
An AWS spokesperson explained that by simply defining prompts and a list of tools, the agent can independently formulate plans and invoke necessary tools to complete tasks. They added that developers can easily test in their local environment before seamlessly deploying to the AWS cloud.
Several AWS internal services, including Amazon Q Developer, AWS Glue, and Amazon VPC Reachability Analyzer, already leverage the Strands Agents SDK.
The Strands Agents SDK integrates with various model services such as Amazon Bedrock, Ollama, and LiteLLM.
Ollama enables local execution of open-source LLMs.
LiteLLM unifies over 100 cloud-based LLM Application Programming Interfaces (APIs) through a single interface, empowering developers to harness diverse LLMs without mastering each API’s specific features.
The SDK offers thousands of Model Context Protocol (MCP) servers and 20 pre-built example tools.
Released under the Apache License 2.0, the Strands Agents SDK welcomes contributions from the open-source community through GitHub and Discord for feature enhancements and documentation. Developers can find installation instructions and sample code on the official GitHub page.