Strands Labs
Strands Labs is the experimental arm of Strands Agents - a space for projects that push the boundaries of what AI agents can do. Labs projects explore new domains, validate novel approaches, and move fast. All projects are open source.
While the core Strands Agents SDK provides the foundation for building agents - the agent loop, tool use, model providers, and multi-agent patterns - Labs is where that foundation gets applied to new problem spaces. These are projects that extend agents into areas like physical robotics, world models, agentic benchmarking, harness optimization, and real-time audio. Some Labs projects may eventually graduate into the core SDK or become standalone products; others may remain experimental. The common thread is that they all build on agentic AI in open source and are designed to be used alongside it.
Labs projects are fully functional and published to package repositories, but they move faster and have a wider surface area than the core SDK. Expect more frequent changes, newer integrations, and a focus on enabling research and prototyping alongside production use.
Projects
Section titled “Projects”Control, simulate, and train physical robots with natural language through Strands Agents. One Robot() call returns a MuJoCo simulation or real hardware - same code, same natural-language control - with pluggable vision-language-action policies, a hardware abstraction layer, and a peer-to-peer mesh.
Bring NVIDIA Cosmos to Strands Agents: give your agent eyes that understand physics and hands that generate video, audio, and robot actions on local compute. Ships four model providers (Cosmos 3 Reasoner and Generator, plus the Cosmos-Reason2 edge VLM) and 45 tools spanning reasoning, generation, curation, post-training, and evaluation.
Strands-based agents and harnesses for agentic benchmarks. Includes Simple Strands Agent (SSA), a lean, hackable autonomous-coding harness that pairs frontier LLMs with shell and file-editing tools in isolated Docker environments, delivering state-of-the-art results on SWE-Bench Verified, SWE-Bench Pro, and Terminal Bench 2.
Optimize an LLM agent’s harness through tunable Formulas that dynamically enhance the agent, then improve those Formulas with optimizers based on collected agent rollout trajectories. A PyTorch-style training loop (DataLoader, rollout engine, reward function, optimizer) for evolving agent context.
Python functions that behave like standard functions but are evaluated by AI agents. AI Functions enforce correctness through runtime post-conditions rather than prompt engineering alone, enabling developers to build reliable agentic workflows using familiar programming patterns. Supports async execution, parallel workflows, and composing functions into multi-agent systems.
Python bindings for WebRTC’s audio processing pipeline - echo cancellation, noise suppression, automatic gain control, and voice activity detection - the same algorithms that run in Chrome. Removes an agent’s own speaker output from its mic capture so a voice agent only hears the user, with a working Strands BidiAgent integration.
Contributing
Section titled “Contributing”Have an experimental idea that pushes AI agents forward? Labs is designed for innovation from across the community. Check the contributing guide to get started.