The Scaling Problem

Most SaaS products go through predictable stages:

1.Prototype — proving the idea works

2.Early users — handling real traffic for the first time

3.Growth — scaling to hundreds or thousands of users

4.Production — high availability, compliance, performance SLAs

At each stage, your AWS infrastructure needs change dramatically. Lambda memory, database instance sizes, VPC configurations, timeout settings, rate limits, and data safety policies all need to evolve.

Traditionally, this means manually updating CloudFormation templates, CDK constructs, or Terraform configs at each stage. It's error-prone, time-consuming, and often requires infrastructure expertise that early-stage teams don't have.

Presets: One Word, Entire Stack

VentureKit's preset system encapsulates all infrastructure decisions for each stage into a single configuration:

typescript

envs: {
  local: { preset: 'free' },
  dev: { preset: 'nano' },
  staging: { preset: 'micro' },
  prod: { preset: 'medium' },
}

Each preset configures your entire stack:

free — Explore ($0)

Local development server with Docker Compose

Postgres, Redis, MinIO running locally

Hot reload, all packages available

No AWS account needed

nano — Prototype (~$5–15/mo AWS)

128 MB Lambda functions

10 second timeouts

10 req/s rate limit

No VPC (minimal cost)

Relaxed data safety (no encryption at rest)

micro — Development (~$30–80/mo AWS)

256 MB Lambda functions

10 second timeouts

50 req/s rate limit

VPC enabled (network isolation)

Standard data safety (encryption at rest, automated backups)

medium — Production (~$100–300/mo AWS)

512 MB Lambda functions

15 second timeouts

100 req/s rate limit

VPC enabled with multi-AZ

Strict data safety (encryption, PITR, enhanced monitoring)

large — High-Performance (~$500+/mo AWS)

1024 MB Lambda functions

30 second timeouts

500 req/s rate limit

VPC with multi-AZ and NAT Gateway

Strict data safety with compliance features

How Presets Work

When you run vk deploy, VentureKit's resolver:

1.Reads your active preset for the target environment

2.Applies preset defaults to every infrastructure intent

3.Deep-merges any per-environment overrides you've specified

4.Generates CDK constructs with the resolved configuration

The key insight is that presets aren't just Lambda configuration — they cascade through your entire stack. A medium preset doesn't just give you bigger Lambdas. It also:

Upgrades your RDS instance to a production-grade class

Enables multi-AZ for database high availability

Turns on Point-in-Time Recovery for backups

Configures enhanced monitoring and CloudWatch alarms

Sets up VPC flow logs for security auditing

Enables encryption at rest for all data stores

Override Anything

Presets are sensible defaults, not constraints. Override any setting by adding the relevant section directly alongside the preset in your environment config:

typescript

envs: {
  prod: {
    preset: 'medium',
    lambda: { memoryMb: 1024, timeoutSec: 20 },
    api: { throttleRateLimit: 200 },
  },
}

Overrides are flat — you specify the category (lambda, api, vpc, observability) at the same level as preset. VentureKit deep-merges your values with the preset defaults, so you only need to specify what's different. Every other setting keeps its preset value.

This gives you the medium preset's database, VPC, and security settings, but with custom Lambda memory and rate limits.

Cost Transparency

All costs shown in VentureKit's documentation are estimated AWS infrastructure costs, not VentureKit pricing. VentureKit itself is open source under Apache 2.0.

The estimates help you plan: start with free to explore the framework, deploy with nano when you're ready to go live, and scale to medium or large as your user base grows. Your code stays exactly the same — only the preset changes.

Getting Started

Try the free preset right now:

bash

npx @venturekit/cli init my-saas
cd my-saas
vk dev

You'll have a full local development environment in under 2 minutes, with Postgres, Redis, and S3-compatible storage — all running on your machine with hot reload.

From Free to Enterprise: How Environment Presets Let You Scale SaaS on AWS Without Refactoring