AI News Feed
Market watch
FAA Rolls Out $875 Million AI Airspace System, Starting at Three Washington-Area AirportsGoogle Confirms Gemini Models Hacked Three Companies in May 2026 TestAmazon Blocks Meta’s AI Agent Muse From Shopping on Amazon.comX to Notify Users When Governments Force Limits on Their PostsKDE Nears 30th Anniversary With Contributors and Commits Trending UpEli Lilly CEO says 700,000 seniors started GLP-1s under Medicare, 70% on Lilly drugs, as Foundayo gains among oral patients
FAA Rolls Out $875 Million AI Airspace System, Starting at Three Washington-Area AirportsGoogle Confirms Gemini Models Hacked Three Companies in May 2026 TestAmazon Blocks Meta’s AI Agent Muse From Shopping on Amazon.comX to Notify Users When Governments Force Limits on Their PostsKDE Nears 30th Anniversary With Contributors and Commits Trending UpEli Lilly CEO says 700,000 seniors started GLP-1s under Medicare, 70% on Lilly drugs, as Foundayo gains among oral patients
Products & Applications

AWS Debuts Strands Harness, an Open-Source AI Agent for Any Environment

AWS has launched Strands Harness, an open-source AI agent designed to move developers from local prototypes to scalable cloud deployments, with support for multiple model providers and claims of better token efficiency.

AI News Feed · 2026-09-22 02:48

According to SiliconANGLE, AWS detailed the launch in a blog post. AWS said many developers have already built local AI agent prototypes using tools such as Anthropic's Claude Code and OpenAI's Codex because those environments worked on their own machines. When developers try to transition those prototypes to scalable cloud infrastructure, however, they often encounter problems.

Strands Harness is built on the Strands Harness software development kit, which AWS describes as an open-source framework for managing and running multi-agent patterns. The agent supports frontier models from Anthropic, OpenAI, Amazon Bedrock and Google. Developers who want to keep their agents hosted locally can also direct Strands Harness to an Ollama model running on their own machine, AWS said.

AWS said Strands Harness ships with essential components including read, write and edit, shell and web search capabilities. Instead of requiring developers to design bespoke tools for each agentic task, the agent relies on tools that the underlying model already knows how to use.

The agent also includes context management features. AWS said it can offload tool results to separate files and cache reused parts of requests, reducing processing time and token consumption. Strands Harness maintains long-term memory across runs and can resume previous conversations through session IDs. It includes a built-in helper agent that can be delegated open-ended subtasks through an automated checklist. Developers can upload Agent Skills and integrate external tools such as Model Context Protocol servers.

AWS reported that Strands Harness performed strongly on a range of industry benchmarks. According to the company, it was 26% more efficient than agents built on other frameworks when using the same underlying model. AWS also said Strands Harness had greater token efficiency than Claude Code and Codex and performed better than those agents on several other benchmarks. In one test using Anthropic's Fable 5 model, AWS said Strands Harness cost 77% less than Claude Code on the same tasks while achieving a higher overall score on the Terminal Bench 2.1 benchmark.

AWS identified rapid prototyping of new AI applications as the primary use case. One AWS team used Strands Harness to develop the Strands CLI, a command-line interface intended to let developers and people without coding skills prototype AI agents using natural-language commands. Users select the underlying model, add the prompts and tools they think the agent needs, and then use an '/export' command to download the underlying harness code as a Python or TypeScript file.

Strands Harness is available for download from GitHub, AWS said. It can also be installed through standard package managers using pip install strands-agents-harness or npm install strands-agents-harness.