AWS creates a sandbox for its agent experiments

The New Stack

by Frederic Lardinois

February 23, 2026

AI-Generated Deep Dive Summary

Amazon Web Services (AWS) has launched Strands Labs, a dedicated GitHub organization focused on experimental agentic AI projects. This new initiative aims to separate cutting-edge, experimental work from the production-ready Strands Agents SDK, ensuring the core tool remains stable and reliable for developers. The initial release includes two notable projects: AI Functions and Strands Robots. AI Functions enables developers to define Python functions using natural-language specifications, allowing an AI agent to generate the code at runtime. This innovation is particularly valuable for handling complex tasks like parsing receipts, where varying formats can make deterministic code brittle. By incorporating guardrails, such as preconditions and postconditions, the tool ensures self-correction if the output isn’t accurate. This approach makes it easier for developers to integrate AI-generated functions into their workflows without disrupting existing deterministic logic. On the other hand, Strands Robots bridges large language models (LLMs) with physical hardware through vision-language-action (VLA) models. This project demonstrates how AI can interact with real-world environments, opening up possibilities for integrating LLMs into robotics and IoT applications. Together, these projects highlight AWS’s commitment to advancing AI capabilities while maintaining a clear distinction between experimental and production-ready tools. The creation of Strands Labs underscores the importance of providing a dedicated space for innovation without risking the stability of established technologies. By isolating experimental work, AWS ensures that the core SDK remains reliable for production use cases, while still allowing internal teams and the broader community to iterate on new ideas. This approach not only fosters innovation but also provides developers with access to early-stage projects that could shape the future of AI-driven development. For readers interested in DevOps and cloud technology, Strands Labs represents a significant step forward in AI experimentation. The availability of these projects under a dedicated organization ensures that developers can explore cutting-edge tools without compromising their existing workflows. As AWS continues to push the boundaries of agentic AI, Strands Labs serves as a hub for innovation, offering valuable resources for both internal teams and the wider developer community.

Verticals

devopscloud

Originally published on The New Stack on 2/23/2026