dataqbs
OSS Tool MIT

Floci: the open-source AWS emulator that replaces LocalStack

Local AWS emulator built on Quarkus Native (MIT). A drop-in replacement for LocalStack Community (which starts requiring an auth token in March 2026) — 24 ms cold start, 13 MiB RAM, ~45 AWS services.

Why it matters Lets you run integration tests against S3, DynamoDB, SQS, Lambda, and dozens more without ever hitting real AWS. Zero credentials, reproducible container, cheap to run in any CI pipeline.

Floci: the open-source AWS emulator that replaces LocalStack

What is it?

Floci is a local AWS emulator written in Quarkus Native and released under the MIT license. It cold-starts in 24 ms and idles at ~13 MiB RAM. It covers roughly 45 services — S3, DynamoDB, SQS, SNS, Lambda, ECS, ECR, OpenSearch, MSK/Kafka, Athena, IAM, Step Functions, among others — by pointing your AWS SDKs at the local endpoint http://localhost:4566.

It is positioned as a direct replacement for LocalStack Community Edition, whose sunsetting in March 2026 forces auth tokens and freezes security patches.

When does it fit?

  • Integration tests against AWS services without burning credits or polluting real accounts.
  • Offline dev loop: on a plane, behind a restrictive client network, or when AWS itself is down.
  • CI pipelines that need reproducible endpoints — 24 ms boot fits inside any job.
  • Multi-account isolation for cross-account IAM tests without provisioning real sub-accounts.
  • Migrating from LocalStack Community without paying for the Pro edition or losing the local-endpoint pattern.

Quickstart

docker run --rm -p 4566:4566 floci/floci

Or in docker-compose.yml:

services:
  floci:
    image: floci/floci:latest
    ports: ["4566:4566"]

Point the SDK at the local endpoint:

export AWS_ENDPOINT_URL=http://localhost:4566
export AWS_ACCESS_KEY_ID=dummy
export AWS_SECRET_ACCESS_KEY=dummy

The endpoint convention is identical to LocalStack, so most existing configuration keeps working unchanged.

Limitations to call out

  • Newer or preview AWS services may not be covered — check the official list before promising one to a client.
  • IAM policy evaluation is “real enough” for most test cases but doesn’t replicate every edge case of the AWS authorizer.
  • It is not a pricing simulator — use AWS Pricing Calculator or Cost Explorer for cost modelling.
  • As with any young OSS project, the service list and the license model can evolve. Verify at the start of each engagement.

Why we are publishing it

At dataqbs we evaluate open-source tools that offer structural advantages over the commercial alternative — cost, license, data sovereignty, simulation fidelity. When one passes the internal filter we document it publicly so other teams can reuse the analysis without redoing it from scratch.

Repository: github.com/floci-io/floci · License: MIT

This note is a recommendation based on internal evaluation. Verify license, maturity, and service coverage before adopting it in production.