Use cases

Why teams use Simfra

A local AWS environment that starts in seconds, costs nothing to run, and behaves like the real thing.

CI/CD testing

Run your full infrastructure test suite against a real AWS-compatible environment in your CI pipeline. No shared AWS account, no credential management, no cleanup scripts.

  • Spin up a fresh environment per pipeline run in seconds
  • Test Terraform modules with plan/apply/destroy cycles
  • Validate IAM policies, resource policies, and cross-account access before merging
  • No flaky tests from shared state or rate limits
  • Run hundreds of parallel pipelines without AWS costs

GitHub Actions

services:
  simfra:
    image: ghcr.io/simfra-dev/simfra:latest
    ports:
      - "4599:4599"

steps:
  - run: terraform apply -auto-approve
    env:
      AWS_ENDPOINT_URL: http://localhost:4599

Local development

Develop against a full AWS environment running on your laptop. No internet required, no shared account conflicts, instant feedback.

  • All 88 services on one port, one container
  • SDKs and CLI tools work unchanged with a single endpoint override
  • Docker-backed Lambda, RDS, ElastiCache, and EKS for realistic development
  • Persist state across restarts with SQLite
  • No risk of accidentally modifying production resources

terminal

$ docker run -p 4599:4599 \
    -e SIMFRA_DOCKER=true \
    -e SIMFRA_DATA_DIR=/data \
    -v simfra-data:/data \
    ghcr.io/simfra-dev/simfra:latest

$ export AWS_ENDPOINT_URL=http://localhost:4599
$ aws sqs create-queue --queue-name orders
$ aws lambda invoke --function-name process-order ...

Rapid prototyping with AI agents

Give AI coding agents access to a full AWS environment with zero wait time, zero cost, and zero risk. Agents can create, test, and iterate on infrastructure and application code against real AWS APIs instantly.

  • No waiting for CloudFormation stacks or provisioning delays
  • Agents can apply and destroy infrastructure thousands of times without cost
  • No risk of runaway costs from agent errors or loops
  • No IAM credentials to leak or manage
  • Full AWS API fidelity means agent-generated code works in production

agent workflow

# Agent iterates freely against local AWS
$ terraform apply     # instant, free
$ terraform destroy   # instant, free
$ terraform apply     # try a different approach

# When ready, same code deploys to real AWS
$ export AWS_ENDPOINT_URL=  # unset
$ terraform apply     # production

Proof of concept work

Validate architectural decisions before committing to AWS spend. Test multi-service architectures, cross-account patterns, and complex integrations locally.

  • Prototype event-driven architectures with real SNS/SQS/EventBridge delivery
  • Test Step Functions workflows against live Lambda and DynamoDB
  • Validate cross-account access patterns with real IAM evaluation
  • Build and test CI/CD pipelines with CodePipeline, CodeBuild, and CodeDeploy
  • Share reproducible demos with your team via Docker Compose

docker-compose.yml

services:
  simfra:
    image: ghcr.io/simfra-dev/simfra:latest
    ports:
      - "4599:4599"
    environment:
      SIMFRA_DOCKER: true
      SIMFRA_DATA_DIR: /data
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      - simfra-data:/data

Reducing AWS development costs

Stop paying for development and staging environments that sit idle or accumulate forgotten resources. Simfra runs on hardware you already have.

  • Eliminate per-developer AWS sandbox accounts
  • No forgotten EC2 instances, RDS databases, or NAT gateways running overnight
  • No data transfer charges for local development
  • No CloudWatch, CloudTrail, or Config costs from dev activity
  • Run the same infrastructure code against Simfra locally and AWS in production

User sandboxes

Give every developer, contractor, or team their own isolated AWS environment. No cross-contamination, no shared resource conflicts, no cleanup coordination.

  • Multi-account support with full IAM isolation between users
  • Each sandbox starts clean or from a bootstrapped baseline
  • No organizational overhead of managing AWS accounts and SSO
  • Sandboxes are ephemeral and disposable by default
  • Admin API for programmatic sandbox creation and teardown

terminal

# Create isolated sandboxes via admin API
$ curl -X POST http://localhost:4599/_simfra/accounts \
    -H "Authorization: Bearer $ADMIN_TOKEN" \
    -d '{"name": "developer-alice"}'

$ curl -X POST http://localhost:4599/_simfra/accounts \
    -H "Authorization: Bearer $ADMIN_TOKEN" \
    -d '{"name": "developer-bob"}'

Training and education

Teach AWS services and cloud architecture without giving students access to a real AWS account. Every student gets their own isolated environment.

  • Students can experiment freely without cost concerns
  • No risk of accidental resource creation in production accounts
  • Built-in web console for visual exploration of resources
  • IAM policy enforcement teaches real permission models
  • CloudShell provides a browser-based terminal for hands-on exercises

Have a different use case?

We'd love to hear how you're using Simfra.