Home / Blog / AWS Releases a Free Local Emulator of AWS Built Entirely by…
Tech News

AWS Releases a Free Local Emulator of AWS Built Entirely by Agents

AWS released AWS Emulator, a free CC0-licensed local tool covering over 200 services with code written and maintained entirely by autonomous AI agents.

By Dillip Chowdary • Nov 20, 2026 • Source: InfoQ

AWS Releases a Free Local Emulator of AWS Built Entirely by Agents
Image: InfoQ

At a glance

Product
AWS Emulator
License
Creative Commons Zero (CC0)
Service catalog
Over 200 AWS services
Default port
9297
Price
Free
Verified version
1.0.322

Amazon Web Services launched AWS Emulator, a free local implementation of more than 200 cloud services dedicated to the public domain under a Creative Commons Zero license. As detailed in InfoQ's report, the software runs as a single Rust binary generated from 438 public AWS API models and listens locally on port 9297. AWS stated that thousands of autonomous artificial intelligence agents wrote every line of code in the project to support local development and automated agent testing loops.

This report covers the architecture, setup requirements, independent benchmark results, security boundaries, and licensing economics of the new local emulator. It is written for cloud architects, backend engineers, and developers building autonomous coding agents who need offline validation without incurring cloud bills. The analysis breaks down how the agent-driven codebase was generated, where fidelity falls short of production environments, and why practitioner trust remains sharply divided.

01Single Rust binary and local endpoint configuration

The emulator packages mock implementations of 200 services into a standalone binary compiled from public API specifications. Developers redirect standard tools by pointing their terminal at the local socket, overriding credential lookups, and disabling EC2 metadata endpoints. Launch post author Daniel Abib confirmed that existing command line scripts, SDKs, and automated agents run without modifying application logic once the environment variables are exported.

Command
export AWS_ENDPOINT_URL=https://127.0.0.1:9297
export AWS_SHARED_CREDENTIALS_FILE=~/.aws/emulator/9297/emulator_credentials
export AWS_CONFIG_FILE=~/.aws/emulator/9297/config
export AWS_EC2_METADATA_DISABLED=true
export AWS_REGION=us-east-1
unset AWS_ACCESS_KEY_ID AWS_SECRET_ACCESS_KEY AWS_SESSION_TOKEN AWS_PROFILE

The same shell configuration allows autonomous coding agents to provision and query resources locally without cloud latency. In demonstration tests, AWS directed its Kiro agent to create a pay-per-request DynamoDB table named kiro-demo and verify immediate item retrieval through standard CLI calls. The agent dispatched routine table creation, record insertion, and fetch requests directly against the local loopback interface.

Command
aws dynamodb create-table --table-name kiro-demo \
--attribute-definitions AttributeName=id,AttributeType=S \
--key-schema AttributeName=id,KeyType=HASH \
--billing-mode PAY_PER_REQUEST
aws dynamodb put-item --table-name kiro-demo \
--item '{"id": {"S": "item-1"}, "msg": {"S": "hello from kiro"}}'
aws dynamodb get-item --table-name kiro-demo --key '{"id": {"S": "item-1"}}'

AWS chief AI and technology officer Matt Wood explained that fast local execution prevents agent divergence during automated cycles. Wood warned that slow feedback causes coding agents to build elaborate software around mistaken assumptions, while flawed emulation can reinforce invalid behavior. To expose discrepancies early, AWS publishes per-service test pass rates against real cloud infrastructure with every release.

02Three agent teams and automated merge validation

Senior principal engineer Matthew Goldberg originated the codebase by giving the Kiro agent an experimental prompt without expecting functional output. Kiro delivered a working Amazon SQS implementation in fifteen minutes, prompting Goldberg to expand the scope across the remaining cloud catalog over a single weekend. The internal engineering group began running team software against the prototype that following Monday.

  1. Write tests Generate test cases derived from the AWS Command Line Interface.
  2. Iterate tests Execute and refine test suites until all assertions pass against production AWS.
  3. Build implementation Construct an in-memory service implementation that satisfies those verified tests.

Production maintenance now relies on three specialized agent groups operating concurrently across the repository without human code authoring. One agent team maintains test suites, a second discovers implementation defects, and a third writes corrective patches. Every submitted fix must compile cleanly and satisfy automated agent reviews for security, coding standards, and fidelity regressions before merging into mainline.

Swami Sivasubramanian recounted using Kiro to rebuild a complex replication engine that had previously required four months of manual developer effort. After spending four nights supervising unproductive cycles, Sivasubramanian provided the agent with explicit specifications and an executable test environment, reducing total rebuild time to two hours. He concluded that human engineering now centers on designing specifications, tests, and merge gates rather than manual programming.

Advertisement

Tech Pulse Daily

Get tomorrow's pulse first

Join engineers who read Tech Pulse before stand-up. Free, weekday mornings.

03Independent fidelity benchmarks across core cloud services

Community measurements provide the first external baseline comparing the emulator against live cloud environments. AWS Community Builder Kento Ikeda evaluated release version 1.0.322 within GitHub Actions continuous integration pipelines immediately following the public announcement. His findings confirmed that basic API shapes match production, although operational fidelity varies dramatically depending on the specific service and operation invoked.

Service or PlatformFidelity ScoreOperational Pass RatePricing Model
Amazon DynamoDB60% to 70%80% to 100%Free (CC0)
Amazon S360% to 70%80% to 100%Free (CC0)
Amazon SQS60% to 70%80% to 100%Free (CC0)
AWS CloudFormation~20%UnspecifiedFree (CC0)
LocalStack HobbyBaseline referenceCommercial restrictedFree (account required)
LocalStack ProCommercial baselineFull cloud mockFrom $39 per month

Core storage and messaging services achieved respectable operational scores, but deeper lifecycle behaviors revealed clear operational divergence. While DynamoDB matched AWS error messages, newly created tables remained in a creating state for three seconds and rejected concurrent write attempts. Ikeda observed that matching public API schemas guarantees payload formatting but does not replicate internal state transitions or underlying cloud mechanics.

Fidelity scores fluctuate because the software derives strictly from public interface definitions rather than internal service code. Abib urged practitioners to view published scores as progress snapshots rather than guarantees of behavioral equivalence. For complex orchestration, CloudFormation achieved approximately twenty percent fidelity, reinforcing Ikeda's assessment that the emulator serves as an intermediate testing tier between lightweight mocks and cloud sandboxes.

04Security boundaries and omitted identity policy validation

The official documentation explicitly advises developers never to rely on the emulator for security validation or access policy testing. The engine omits AWS Identity and Access Management policy evaluation as well as resource-based policy enforcement entirely. Consequently, requests that production AWS environments strictly deny will execute without restriction when routed through the local service endpoint.

Note: AWS warns developers to keep real cloud credentials completely out of local emulator test environments. Because the binary never denies requests signed with its own internal key, teams must test security configurations and IAM policies on real AWS infrastructure before deploying applications to production environments.

The runtime architecture excludes several critical distributed systems behaviors that engineers encounter in production systems. The binary returns HTTP 501 status codes for unsupported operations, transmits usage telemetry by default, and makes no attempt to simulate request throttling, regional latencies, or service outages. Furthermore, the underlying code repository does not accept external pull requests, leaving maintenance entirely to internal automated agents.

05Competitive licensing pressures and practitioner reception

The CC0 public domain release arrives amidst significant friction across the local cloud emulation ecosystem. Competitor LocalStack recently updated its free Hobby tier to require registered accounts and prohibit commercial use, while pricing commercial seats from 39 dollars monthly. Those licensing shifts pushed several development teams toward alternatives like Floci, making a free AWS release appealing despite its early functional limitations.

Community sentiment remains sharply polarized regarding whether to trust a codebase authored exclusively by artificial intelligence. One Hacker News participant noted that 29,684 commits in a repository under a month old did not inspire confidence, while Reddit commenters criticized pre-production testing against unreviewed code. Conversely, defenders argued that generating services against rigid API specifications is an ideal task for iterative autonomous agents.

AWS verified that the emulator represents a permanent product commitment rather than an ephemeral research demonstration. Because the software carries a CC0 dedication, external maintainers and competing tools can freely inspect or fork the generated Rust codebase. Executive Swami Sivasubramanian stated that AWS will share further technical details regarding its agent-driven development architecture during the re:Invent conference in December.

Frequently asked questions

Does AWS Emulator evaluate IAM permissions and security policies?

No, the emulator completely omits IAM and resource-based policy evaluation, and it never denies a request signed with its own key. AWS instructs developers to keep real credentials out of test jobs and verify security on live infrastructure.

How was the AWS Emulator codebase developed and maintained?

Thousands of autonomous AI agents generated and maintain the entire Rust codebase from AWS's 438 public API models. Three agent teams write tests, identify bugs, and submit fixes through automated merge validation gates without human code review.

What fidelity scores did independent benchmarks measure for core services?

In community testing of version 1.0.322, DynamoDB, S3, and SQS passed 60 to 70 percent of service-level tests and 80 to 100 percent of called operations. CloudFormation achieved approximately 20 percent fidelity.

How does AWS Emulator compare in pricing to LocalStack?

AWS Emulator is free under a CC0 public domain dedication and permits unrestricted commercial use. LocalStack's free Hobby plan requires an account and prohibits commercial usage, with paid commercial seats starting at $39 per month.

Sources

Dillip Chowdary

Author

Dillip Chowdary

Writes Tech Bytes coverage of AI, engineering, and the tools that actually ship. Editor of Tech Pulse Daily.

Related on Tech Bytes

Advertisement

5-min tech signal

Weekday briefing for engineers who skip the noise.

No spam · Unsubscribe anytime

Advertisement

✈️ CareerPilot

Your AI job-search copilot

Match your resume against live Ashby, Greenhouse & Lever openings — fit scores, job-specific resume optimization and email alerts.

Find matching jobs →

Free Tools

Browse all tools →