Science & Tech · AI · 2 hrs ago
AWS releases a local cloud-testing tool written entirely by AI agents
Amazon has released AWS Emulator, a tool developers can run on their own computers to test software.
The project began as a weekend experiment by an AWS team using the Kiro AI agent and a set of tests based on the AWS Command Line Interface.
AI agents wrote all of its code, but the human AWS team has not reviewed it.
Despite its name, the tool is not a full emulator: it provides local versions of some AWS API endpoints, but not every cloud feature.
Developers can test without calling AWS or deploying their software, which can help them find problems before using the real cloud service.
AWS says turning the experiment into a product required further engineering work, and the tool is available on GitHub; Mac and PC users run it in Docker, while Linux users can install packages for their distribution.
Amazon has released an AWS Emulator whose code was written entirely by AI agents.
The tool provides local API endpoints for developers to test software without calling AWS.
It is not a full emulator, and it does not offer all AWS cloud features.
The project began as a weekend experiment using AWS’s Kiro autonomous agent.
AWS engineer Matthew Goldberg said making it a product required further engineering and guardrails.
- Who
- Amazon’s AWS team developed the tool using its Kiro autonomous agent.
- What
- It released a local tool for testing software against replicated AWS API endpoints.
- When
- The article was published on November 16, 2026. The initial version was built in 15 minutes.
- Where
- The code is available on GitHub, and the tool runs locally on a user’s machine or development server.
- Why
- It is intended to help developers test whether software will work on AWS without deploying it to the cloud.
This story does not have two clearly opposing sides.
The AWS team began the project as an experiment inspired by a demonstration of autonomous AI agents building a C compiler.
The agent produced a passable version of the tool.
Goldberg had refined the prompt to produce a version his team could use.
- Code
- Entirely written by AI agents
- Initial build time
- 15 minutes
- Development agent
- Kiro
- License
- CC0
- Availability
- GitHub
- Deployment requirements
- Docker on Mac and PC; RPM or Debian package on Linux
1 source
What people are saying
Turning that into a robust product took more than a weekend.
building something that you feel confident putting in your customers’ hands still requires serious engineering effort.
You use it to test your software without calling AWS, on your own machine, on a development server, or in continuous integration.