How to run
AWS locally.

Hitting real AWS during development is slow, costs money, and needs a network and credentials. Running AWS locally fixes all three: instant iteration, zero spend, and a sandbox you can reset at will. There are three common ways to do it.

Your options at a glance

• Moto — a Python library that mocks AWS calls in-process. Perfect for fast unit tests; AWS-only and ephemeral. (Vyomi vs Moto →)
• LocalStack — a running AWS emulator. Strong AWS coverage; AWS-only, and the deeper features sit behind a paid tier. (Vyomi vs LocalStack →)
• Vyomi — the Multi-Cloud Digital Twin Runtime. Runs AWS plus GCP and Azure on real backends, with native consoles and Terraform export. Best when your stack isn’t AWS-only, or you want a real environment rather than mocks. (See the full landscape →)

The rest of this guide uses Vyomi, because the same steps then work for GCP and Azure too.

Step by step

1Install & start

Install via your package manager and start the appliance:

brew install vyomi-cloud/tap/vyomi   # or: docker run -p 9000:9000 vyomi/appliance
vyomi up

Vyomi now answers AWS API calls at http://localhost:9000.

2Point boto3 at it

The only change to your code is the endpoint URL — credentials can be any placeholder:

import boto3

s3 = boto3.client(
    "s3",
    endpoint_url="http://localhost:9000",
    aws_access_key_id="test",
    aws_secret_access_key="test",
    region_name="us-east-1",
)
s3.create_bucket(Bucket="demo")
s3.put_object(Bucket="demo", Key="hello.txt", Body=b"hi from local AWS")
print([o["Key"] for o in s3.list_objects_v2(Bucket="demo").get("Contents", [])])

3Use the aws CLI

aws --endpoint-url http://localhost:9000 s3 mb s3://demo
aws --endpoint-url http://localhost:9000 s3 ls

4Drive it with Terraform

Point the AWS provider at the local endpoint, then terraform apply as usual. When you’re ready for production, Vyomi can also export your running stack back to HCL.

provider "aws" {
  access_key                  = "test"
  secret_key                  = "test"
  region                      = "us-east-1"
  skip_credentials_validation = true
  skip_requester_charged      = true
  endpoints { s3 = "http://localhost:9000" }
}

Going multi-cloud

Because Vyomi runs GCP and Azure from the same appliance, the identical pattern — change the endpoint, keep the SDK — works for google-cloud-storage and azure-storage-blob too. That’s the difference between a local AWS emulator and a full multi-cloud digital twin.