TLDR: Amazon has officially launched Strands Agents 1.0, an open-source SDK designed to significantly simplify the development and deployment of production-ready multi-agent AI systems. This release introduces new primitives for complex agent orchestration, enhanced scalability, and broad model provider compatibility, building on its successful preview phase.
Amazon has announced the general availability of Strands Agents 1.0, an open-source Software Development Kit (SDK) aimed at making the creation and deployment of AI agents, particularly multi-agent systems, more accessible and robust for developers. The release, detailed on the AWS Open Source Blog on July 15, 2025, marks a pivotal step in enabling complex AI architectures for production environments.
Strands Agents, which saw over 2,000 GitHub stars and 150,000 PyPI downloads since its preview launch in May 2025, is built on a model-driven approach. This allows developers to construct and operate AI agents with minimal code, scaling from simple use cases to intricate production deployments. The core philosophy centers on leveraging the reasoning capabilities of Large Language Models (LLMs) to plan, orchestrate tasks, and reflect on goals.
Version 1.0 brings significant enhancements, particularly for multi-agent applications. It introduces four new intuitive primitives—Agents-as-Tools, Swarm, Graph, and Workflow—that simplify the orchestration of multiple specialized AI agents working collaboratively. These patterns enable agents to delegate tasks, share knowledge, and coordinate actions to tackle problems beyond the scope of single agents. A key addition is support for the Agent-to-Agent (A2A) protocol, an open standard facilitating seamless communication between agents from different platforms.
For production readiness, Strands 1.0 incorporates a new session manager for retrieving agent state from remote datastores and features improved asynchronous support throughout the SDK. The platform boasts extensive flexibility, supporting a wide array of model providers including Amazon Bedrock, Anthropic, Meta, OpenAI, Cohere, Mistral, Stability, Writer, and Baseten, along with local models via Ollama. This model-agnostic approach allows developers to switch providers without altering their core code.
Observability is a cornerstone of Strands Agents, with built-in support for traces, metrics, and logs. Traces provide insights into system prompts, model parameters, input/output messages, token usage, and tool invocations, aiding in understanding agent behavior. Metrics offer performance data such as loop statistics, latency, memory, and CPU usage, crucial for hardware decisions. The SDK also provides native AWS integrations, simplifying deployment to services like Amazon EKS, AWS Lambda, and Amazon EC2.
Also Read:
- AWS to Launch AI Agent Marketplace, Featuring Anthropic and OpenAI Models
- AWS’s Strategic Push for Affordable and Agentic AI: An Interview with Amazon Bedrock Chief Atul Deo
Several major technology and consulting firms, including Accenture, Anthropic, Langfuse, mem0.ai, Meta, PwC, Ragas.io, and Tavily, have contributed to and expressed support for the Strands project. Internally, AWS teams like Amazon Q Developer, AWS Glue, and VPC Reachability Analyzer are already utilizing Strands in their production systems, underscoring its reliability and practical utility in real-world applications.
