AWS has released a free, open-source emulator that lets developers test software against a local stand-in for AWS services, including in continuous integration pipelines. The unusual part is how it was built: AWS says autonomous agents wrote and maintain the code, without engineers writing or reviewing it.
AWS AI Watch analysis
What happened
The emulator runs on a developer’s machine or development server. Developers can use it to test how their applications interact with AWS services without deploying to AWS or using cloud resources; AWS charges no additional fee for the emulator itself. The source code and tests are available on GitHub under a CC0 public-domain dedication.
Help Net Security reports that the system is developed by three teams of agents: one creates tests from public AWS API specifications and checks them against real AWS, another compares those tests with the emulator, and a third works on fixes. Changes must pass build checks and reviews by other agents for coding standards, security and fidelity. AWS says the test collection has grown to tens of thousands, and that hundreds of agents work on the code simultaneously.
Why it matters
Local cloud testing can make it quicker and cheaper to run integration tests or prototype applications before deployment. Here, the development process is news too: AWS says agents produce the code, while other agents check changes against real services and run review steps. That is a concrete example of a company using agents to maintain a substantial software tool, not just a demo that writes a function and calls it a day.
There are important boundaries. AWS says the emulator does not evaluate IAM or resource-based policies as the real service does, so it should not be used to test security. Resources disappear when it stops, fidelity varies by service, and AWS has not tested it for long-running use or load testing. Anonymous usage and error data is sent by default, though users can turn that reporting off.
Our read
For development and CI, a free local emulator is an appealing way to test ordinary service interactions without making every test depend on a live cloud environment. The agent-built process is striking, but it does not make the emulator a faithful substitute for AWS in every case. Use it for the local development work AWS describes, not as a security test or production stand-in. That distinction is less glamorous than “AI built the cloud”, and rather more useful.
What to watch
- How fidelity scores change as AWS adds tests and implements more service behaviour.
- Whether developers find the emulator useful in local and CI workflows.
- Whether AWS publishes more detail about agent review and the limits of its security checks.
Discussion spark: Would you trust an AI-maintained emulator for everyday development and CI while reserving security testing for real AWS, or is that too sharp a boundary?
Sources and evidence
- Source update (12 November 2026, 09:30 UTC)
not affiliated with or endorsed by Amazon Web Services (AWS)