AWS Emulator: An open-source tool for testing applications locally without cloud costs
Amazon Web Services (AWS) has announced the launch of AWS Emulator, an open-source local emulator developed in Rust that allows developers to run and test their applications against most AWS services directly from their personal machines. It uses the same APIs, SDKs, and AWS CLI tools without the need to deploy resources in the actual cloud environment or pay additional service fees.
Mechanism and direct integration with the development environment
The emulator operates via a secure local endpoint (HTTPS) that listens by default on the local address (127.0.0.1), ensuring that traffic remains contained within the developer's machine. When launched, the emulator creates temporary credentials and a unified dummy account with the ID (000000000000), using access keys that begin with EMULATOR to distinguish them from real AWS keys and avoid confusion.
The emulator is available as ready-to-use packages for Linux operating systems in RPM and Debian formats, while it can be run via Docker on macOS and Windows. Thanks to its construction based on AWS's open Smithy models, the tool provides responses and errors that match the structural format of the real cloud environment.
Support for AI agents and embedded mode
The emulator is designed to suit modern programming methods, especially when using AI coding assistants like Kiro or Claude Code. A digital agent can create simulated cloud resources such as DynamoDB tables or S3 buckets, test the application, iterate, and modify code immediately without waiting for cloud deployments to complete across different regions.
In addition to running as a standalone server, AWS provides an Embedded Mode for Rust developers through the aws-emulator-embedded library. This mode allows the emulator to be embedded within the test process itself (In-Process) and linked directly to the SDK client. Each instance maintains its own state space, allowing for the execution of hundreds of parallel unit tests without resource interference.
Measuring fidelity and current limitations
To ensure the tool's reliability, AWS relies on a set of Fidelity Tests that compare the emulator's responses with the actual cloud environment in terms of status codes, structural responses, and error messages. The results of these tests are published publicly and updated in a dedicated GitHub repository to show the coverage percentage for each service.
However, AWS has identified some important limitations that developers should consider:
- Security policy evaluation: The emulator does not evaluate Identity and Access Management (IAM) policies with the same rigor and precision as AWS, so requests that might be rejected in the cloud could succeed locally.
- Single-account scope: The emulator operates on a single local account system without support for features that require multi-account or multi-region operations.
- AI-driven development: Large parts of the emulator were built and maintained using AI agents via the Amazon Bedrock AgentCore environment, with independent automated security reviews conducted before deployment.
Why does this announcement matter to developers?
AWS Emulator solves a major hurdle in the cloud application development lifecycle. Previously, the testing process always required an internet connection, the creation of test accounts, the incurring of financial costs associated with resource usage, and the slowness resulting from waiting for cloud deployments to complete. This emulator gives software teams the ability to execute integration tests at high speed and block unintended costs during early development stages, while emphasizing the necessity of performing final security and protection policy tests on the real cloud before launching applications in a production environment.