AWS Emulator Delivers Open-Source Local Testing for Over 200 AWS Services

AWS released the AWS Emulator on November 11, 2026, making a free open-source local emulator for more than 200 AWS services generally available. The tool runs on developers’ machines, dev servers, or in continuous integration environments and supports the same request and response formats as the real cloud APIs.

The release addresses repeated testing demands created by AI coding agents that require frequent API validation during development loops. It covers 438 public AWS API models and returns identical errors for implemented operations, allowing code to run without network access or an AWS account.

How the AWS Emulator Handles API Requests and State

The AWS Emulator operates as a lightweight Rust-based HTTPS server that listens on port 9297 by default. Developers redirect the AWS CLI, SDKs, or AI coding agents to the local endpoint by setting environment variables, leaving application code unchanged. The server maintains in-memory state for each instance and supplies its own credentials and configuration.

Installation options include RPM and Debian packages for Linux, Docker images for macOS and Windows, and source builds. An embedded mode supports direct use inside Rust test suites. The package also ships as a skill for AI coding agents and requires AWS CLI version 2.35.0 or later. All operations run offline with no external dependencies.

Users receive fidelity scores and verified test results with each release. Unsupported operations return HTTP 501 Not Implemented. The emulator does not simulate IAM policy enforcement, cross-account behavior, throttling, or region-specific features, so it serves only as a complement to later cloud testing.

Autonomous AI Agents Built and Maintain the Entire Codebase

Autonomous agents constructed every line of the AWS Emulator using Amazon Bedrock AgentCore, CodePipeline, and CodeCommit. Three agent teams generated the code, created tests from API models, verified behavior against live AWS endpoints, and applied fixes through automated pipelines. Senior Principal Engineer Matt Goldberg stated that the team has neither written nor reviewed any of the source code.

The project began as a weekend experiment when Goldberg prompted the Kiro coding agent to emulate Amazon SQS. A working version appeared in roughly fifteen minutes, prompting expansion with added guardrails for robustness and broader coverage. Agents continue to update the emulator by monitoring new API models and regenerating tests without human intervention.

The repository on GitHub at aws/aws-emulator publishes the full commit history, an evaluation suite, and a public backlog labeled “things-that-do-not-work.” The code carries a CC0 1.0 public domain dedication, with the vendor directory under Apache 2.0. No pull requests are accepted; agents manage all changes.

Practical Use Cases and Planned Updates

Teams use the AWS Emulator for local development, CI/CD pipelines, and rapid iteration with AI coding agents. These workflows avoid repeated cloud deployments and associated costs while maintaining API compatibility. The tool supports SaaS development and cloud testing scenarios that previously relied on third-party emulators with narrower coverage.

AWS positions the emulator as an official, high-fidelity alternative that stays synchronized with service updates through automated model ingestion. It does not replace final validation on real AWS infrastructure, particularly for security or performance testing.

Further details on the multi-agent architecture, fidelity metrics, and quality gates will appear at re:Invent later in November 2026. Ongoing agent-driven releases are expected to expand service coverage and raise fidelity scores over time.