AWS Launches Strands Labs to Give Developers a Sandbox for Autonomous AI
Strands Labs lets developers experiment safely with robotics, simulations, and AI functions, accelerating learning, prototyping, and innovation in agentic AI
Last summer, Amazon Web Services (AWS) introduced Strands Agents, an open-source SDK that lets developers build and run AI agents with just a few lines of code. The project has since gained momentum, exceeding 14 million downloads. And now, the company is introducing its next step: Strands Labs, a new GitHub organization aimed at helping developers explore and experiment with cutting-edge agentic AI techniques.
At launch, Strands Labs is available for three specific projects: robots, robots sim, and AI functions.
“Strands Labs is more about exploring the frontier of agentic experiences than building production applications. It’s about looking at what’s next for agents, in collaboration with our developer community,” Clare Liguori, AWS’s senior principal engineer, tells The AI Economy in an email. “I’m personally excited to see what developers build on top of these projects and to learn from them.”
She explains that AWS selected robotics and AI functions because the company considers these areas the ones that most clearly reveal what changes when software becomes agentic. “We think these frontiers should be explored in the open—democratizing innovation, learning, and growth in agentic development,” Liguori says.
Robots
AWS wants to understand how AI-powered agents can be extended to the edge and the physical world. Obviously, any innovation is beneficial to the company, as it could improve the robots it uses for deliveries and warehouse operations, and develop intelligent robots for the home. Physical AI agents powered by Strands Agents will be able to control robots by connecting AI capabilities to physical sensors and hardware.
Robots Sim
Robots Sim refers to a simulation in which robots are put through tasks in environments modeling physics, sensors, and real-world constraints. Through Strands Labs, developers will be able to use Strands Agent to connect agentic robots to these simulations, facilitate rapid prototyping and algorithm development, all without requiring any physical robotic hardware.
AI Functions
The last project supported by Strands Labs is AI Functions, which writes Python functions using natural-language specifications rather than code. AWS claims that Strands Agents will eliminate the need for developers to inspect AI-generated code when reasoning about function behavior using intent specifications.
Although Strands Labs currently hosts only these three projects, AWS plans to add more in the future.
So what was the thinking behind AWS launching Strands Labs as a standalone GitHub organization? Simply put, the company wanted to promote innovation through experimentation. This could be challenging within the same organization as the Strands Agents SDK, which is stable, well-documented, and already public. By keeping it in its own account, AWS can give developers the flexibility to try new things without fear of breaking something already in production.
“The SDK has become a critical dependency for a lot of teams…so Strands Labs gives us, and the broader community, a dedicated space to experiment boldly,” Liguroi confirms. Depending on the experiments run in Strands Labs, don’t be surprised if some eventually make their way into the official Strands SDK in the future.