Amazon Web Services Inc. launched Strands Harness on September 21, 2026. The new open-source AI agent aims to solve the problem of scaling artificial intelligence agents to cloud environments. AWS explained the release in a blog post today.
Many developers build prototypes of AI agents locally using tools like Anthropic PBC’s Claude Code and OpenAI Group PBC’s Codex. These environments work smoothly on local machines. Developers run into problems when transitioning from a local setup to a scalable cloud environment.
Strands Harness provides a complete, ready-to-use agent. It can run locally or in any cloud. Supported platforms include AWS, Google Cloud, Microsoft Azure, Modal, and Cloudflare.
The software is built atop the Strands Harness software development kit. This open-source framework was developed by AWS for managing and running multi-agent patterns. The foundation runs atop frontier models from Anthropic, OpenAI, Amazon Bedrock, and Google. Developers can also point Strands Harness to an Ollama model hosted locally.
Strands Harness includes read, write, edit, shell, and web search capabilities out of the box. The system relies on tools the underlying model already knows how to use instead of requiring bespoke tools for each agentic task.
The model manages its own context window by offloading tool results to separate files and caching reused request parts. This reduces processing times and token consumption. Strands Harness also maintains a long-term memory across runs to resume previous conversations through session IDs. A built-in helper agent handles open-ended subtasks through an automated checklist, and developers can upload Agent Skills or integrate Model Context Protocol servers.
AWS reported that Strands Harness performed strongly on industry benchmarks. The agent was 26% more efficient than agents built on other frameworks using the same underlying model. It also achieved greater token efficiency compared to Claude Code and Codex.
In a test using Anthropic’s Fable 5 model, Strands Harness cost 77% less than Claude Code on identical tasks. It also achieved a higher overall score on the Terminal Bench 2.1 benchmark.
The primary use case is rapid prototyping of new AI applications. An AWS team used Strands Harness to build the Strands CLI command-line interface. This tool lets developers and non-coders prototype AI agents using natural language commands by selecting a model, adding prompts and tools, and using an export command to download the harness code as a Python or TypeScript file.
Strands Harness is available to download from GitHub. Users can install it directly via standard package managers using pip install strands-agents-harness or npm install strands-agents-harness.
