TL;DR
Get the latest gadgets delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
Fakecloud, a local AWS cloud emulator for integration testing, has been released with coverage of 105 AWS services and first-party test SDKs in six languages. According to its documentation, it runs as a standalone binary or Docker image with no account or token required, and reports 248,557 of 248,557 Smithy-generated test variants passing. Independent verification of its conformance claims is not yet available.
A new open-source tool called Fakecloud, described by its developers as a local AWS cloud emulator for integration tests, is now available as a standalone binary or Docker image. According to the project’s official documentation, the emulator covers 105 AWS services and ships with first-party test SDKs for TypeScript, Python, Go, PHP, Java, and Rust, letting applications use their normal AWS clients against localhost while tests inspect the resulting state through dedicated introspection endpoints.
The project’s website states that Fakecloud allows developers to run their production AWS SDK code, the AWS CLI, and infrastructure-as-code tools against a local endpoint without an account, auth token, or paid plan. Once the application has exercised the normal AWS APIs, tests can use the Fakecloud SDKs — which wrap a set of /_fakecloud/* endpoints — to reset state, make assertions, and manually trigger async processors such as email delivery or message fanout.
According to the project’s own figures, the emulator implements 3,932 API operations across 105 services, including S3, SQS, SNS, EventBridge, Lambda, DynamoDB, EC2, IAM, Cognito, Step Functions, API Gateway, and Bedrock. The project reports that 248,557 of 248,557 Smithy-model-generated test variants pass on every commit, a figure it describes as “true 100% conformance,” backed by end-to-end tests against the official AWS SDKs. These are vendor claims; no independent benchmark has been published.
The documentation also lists more than 30 service-to-service integrations said to be wired up, including S3 event notifications, SNS fanout, EventBridge targets, DynamoDB Streams, CloudWatch Logs subscriptions, Cognito triggers, and API Gateway to Lambda invocations. The comparison table on the project site states that Fakecloud starts in roughly 300 milliseconds, uses about 10 MiB of idle memory, and ships as a roughly 19 MB binary — figures that contrast with the roughly 1 GB Docker image it attributes to LocalStack Community.
What This Means for Local Testing Workflows
The tool targets a persistent pain point in cloud-native development: integration tests that depend on real AWS services are slow, costly, and can have side effects, while traditional mocks do not reproduce how AWS services actually behave or interact with each other. An emulator that accepts standard AWS SDK traffic lets teams test the code they ship rather than a parallel mock implementation.
The project also stakes out a licensing and access position. Fakecloud is released under AGPL-3.0, while LocalStack Community — the most widely used alternative — is proprietary and, according to Fakecloud’s comparison, requires an account and token even for core development flows, with several service families such as RDS, ElastiCache, API Gateway v2, and Cognito User Pools reserved for paid plans. Fakecloud lists emulated implementations for each of these in its free tier. The comparison figures come from Fakecloud itself and have not been independently verified.
The availability of test assertion SDKs in six languages is a further differentiator, because emulated state inspection is often the weakest part of local AWS testing — teams otherwise resort to polling loops, fixtures, or private test hooks to verify what their application actually did.
How Fakecloud Compares to LocalStack
: “LocalStack has been the dominant local AWS emulator for years, offering a Community edition with a subset of services and paid tiers for broader coverage and features like persisted state. Fakecloud’s comparison table claims advantages in startup time, memory footprint, install size, and service availability on the free tier, and notes that its test SDKs cover six languages versus LocalStack’s Python and Java offerings.
Examples from the table include 110 SES v2 operations with real inbound receipt rule actions (compared with stored-but-never-executed behavior attributed to LocalStack), 163 RDS operations with PostgreSQL, MySQL, and MariaDB support via Docker, 214 Bedrock operations across four APIs where LocalStack offers no equivalent, and full Cognito User Pools coverage with 122 operations. These characterizations of LocalStack reflect Fakecloud’s own documentation as of its launch and may not track LocalStack’s current feature set.
Fakecloud’s quick-start flow is a one-line install script, a single fakecloud command, and pointing the standard AWS CLI at http://localhost:4566 with any dummy credentials — the same default port LocalStack uses, easing migration between the two.
“fakecloud gives you a local AWS environment that behaves like infrastructure, not a mock. Your app uses the regular AWS SDK, CLI, and IaC tools.”
— Fakecloud project documentation
Claims Awaiting Independent Verification
All performance figures, service counts, and conformance results cited in this article come from Fakecloud’s own documentation and have not been independently tested. The 100% conformance claim refers to Smithy-model-generated API shape tests for the 3,932 implemented operations; it does not necessarily mean behavioral parity with real AWS in every scenario, and the project itself acknowledges the service mix is “tuned for local development and integration tests” rather than complete AWS replication.
The comparison table’s characterizations of LocalStack Community — including its service availability, pricing boundaries, and runtime behavior — reflect Fakecloud’s marketing and may not match LocalStack’s current release. It is not yet clear how Fakecloud performs under sustained load, how faithfully it reproduces edge-case AWS behavior such as eventual consistency or throttling, or how quickly new AWS API versions are tracked. The project’s governance, funding model, and long-term maintenance commitments are also not described in the available material.
Adoption Questions and Roadmap Unknowns
Developers can evaluate Fakecloud today via its install script, Docker image, and SDK packages on npm, PyPI, pkg.go.dev, and crates.io. The natural next steps for the project are real-world adoption reports, third-party benchmarks against LocalStack and Moto, and continued expansion of the service list — the site already labels its EKS support as “batch 1,” implying further control-plane work is planned.
Teams considering a migration will be watching whether the emulator’s conformance suite keeps pace with AWS API changes, and whether the AGPL-3.0 license affects adoption inside companies with restrictive open-source policies. No public roadmap dates have been announced.
Key Questions
What is Fakecloud?
Fakecloud is a local AWS cloud emulator for integration testing. Applications connect their normal AWS SDKs, CLI, and IaC tools to a local endpoint, while tests use companion Fakecloud SDKs to inspect state, reset environments, and control async behavior.
Is Fakecloud free and open source?
According to the project, Fakecloud is licensed under AGPL-3.0 and requires no account, auth token, or paid plan. It runs as a standalone binary or Docker image.
How does Fakecloud differ from LocalStack?
Per Fakecloud’s own comparison, it offers more services on its free tier, faster startup, a smaller footprint, and test SDKs in six languages, while LocalStack Community is proprietary and requires an account and token. These claims come from Fakecloud and have not been independently verified.
Which AWS services does Fakecloud emulate?
The project lists 105 services, including S3, SQS, SNS, EventBridge, Lambda, DynamoDB, EC2, IAM, Cognito, Step Functions, API Gateway, RDS, ElastiCache, EKS, and Bedrock, with 3,932 implemented API operations.
Does 100% conformance mean it behaves exactly like AWS?
No. The claim refers to passing 248,557 Smithy-model-generated API test variants for implemented operations. Behavioral parity with real AWS in every edge case — such as eventual consistency or throttling — is not guaranteed by that metric.
Source: hn
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
