⚡ LaunchStack

Production-ready project scaffolding for Node.js developers.

LaunchStack is an interactive CLI that generates fully structured, dependency-ready projects in seconds. It detects your environment, asks the right questions (or lets AI answer them for you), and outputs a project wired with the exact stack you chose — with no pinned outdated dependencies.

Why LaunchStack?

Starting new projects usually means:

• manually creating folders
• installing and configuring dependencies
• wiring databases and authentication
• setting up Docker and environment variables
• writing repetitive boilerplate

LaunchStack removes that friction by generating a production-ready project structure in seconds. Instead of spending time on setup, developers can start building immediately.

Installation

# Run without installing
npx launchstack

# Or install globally
npm install -g launchstack

Requires Node.js v20 or higher.

Quick Start

npx launchstack init my-app

LaunchStack will guide you through:

1. Selecting a project type
2. Choosing your stack (framework, database, ORM, auth, modules)
3. Optionally describing your app in plain English for AI-assisted configuration
4. Installing the latest stable dependencies automatically

How LaunchStack Works

Developer
   │
   ▼
launchstack init
   │
   ▼
Environment Detection
(Node • Docker • Redis • Git • Ollama)
   │
   ▼
Interactive Wizard or AI Description
   │
   ▼
Stack Configuration
(Framework • DB • ORM • Auth • Modules)
   │
   ▼
Template Generator
   │
   ▼
Dependency Installer
(latest stable versions)
   │
   ▼
Production-ready project

Commands

launchstack init [name]

Launch the full interactive wizard. Optionally pass a project name.

launchstack init
launchstack init my-saas-app
launchstack init my-api --no-install

launchstack create <type>

Quickly scaffold a project by type without going through the full wizard.

launchstack create backend
launchstack create frontend
launchstack create fullstack
launchstack create microservice
launchstack create worker
launchstack create cli

launchstack add <module>

Add a module to an existing LaunchStack project.

launchstack add redis
launchstack add logging
launchstack add email
launchstack add rate-limit
launchstack add queue
launchstack add file-storage
launchstack add payments

Run this from inside a project created with LaunchStack (requires launchstack.json).

launchstack doctor

Check your environment and scan all template dependencies for freshness.

launchstack doctor

Output example:

✔ Node detected
✔ npm detected
✔ Git detected
✔ Docker detected
✖ Redis not detected
✖ Ollama not detected

Checking template dependencies...

  ✔  express v5.0.0
  ✔  fastify v5.2.0
  ✔  prisma v6.3.1
  ⚠  mongoose — last updated >12 months ago

launchstack templates update

Pull the latest templates from the remote registry.

launchstack templates update

launchstack templates list

List all available template categories and variants.

launchstack templates list

Output example:

Available templates:

  backend/
    express
    fastify
    nest
    django

  frontend/
    next
    react-vite
    vue
    svelte
    alpine
    static

  modules/
    logging
    rate-limit
    email
    redis
    queue
    payments

  docker/
    node
    python

Project Types

Supported Stack

Backend Frameworks

Node: Express, Fastify, NestJS Python: Django, Flask

Frontend Frameworks

• Next.js
• React (Vite)
• Vue (Vite)
• Svelte
• Alpine.js (for lightweight interactive UIs)
• Static HTML

Databases

PostgreSQL, MySQL, MongoDB, SQLite

ORMs

Prisma, Drizzle, Sequelize, Mongoose

Authentication

JWT, OAuth (Passport)

Optional Modules

Rate limiting, Logging (Winston), Email (Nodemailer), Redis (ioredis), Background Jobs (BullMQ), File Storage (AWS S3)

Payments

Stripe, PayPal, Paystack, Flutterwave

AI-Assisted Mode

LaunchStack can automatically analyze your app description and suggest a stack configuration. To enable AI mode, you must have at least one supported AI provider available.

Enable AI Mode

LaunchStack supports multiple AI providers. It will automatically use the first available provider in this order:

1. Ollama (recommended) – local AI, no API key required
2. GitHub Models – requires a GitHub personal access token (free tier available)
3. Manual prompts – fallback when no AI provider is available

Option 1: Install Ollama (Recommended)

Ollama allows LaunchStack to run AI locally without any API keys.

Install Ollama:

Mac (Homebrew):

brew install ollama

Mac (MacPorts):

sudo port install ollama

Linux:

curl -fsSL https://ollama.com/install.sh | sh

Windows: Download and run the installer from ollama.com

Start the service:

ollama serve

Pull a model:

ollama run llama3

After installation, LaunchStack will automatically detect Ollama during the environment check.

Option 2: Use GitHub AI Models

LaunchStack can also use GitHub's AI models if you provide a personal access token.

Create a token here: https://github.com/settings/tokens

Then set the token in your shell:

Mac/Linux:

export GITHUB_TOKEN=your_token_here

To make it permanent:

echo 'export GITHUB_TOKEN=your_token_here' >> ~/.zshrc

Windows (PowerShell):

$env:GITHUB_TOKEN="your_token_here"

To make it permanent on Windows, add it via System Properties → Environment Variables.

Restart your terminal and LaunchStack will detect the token automatically.

If no AI provider is detected, LaunchStack will fall back to the manual interactive prompts.

Example:

Describe your app in one sentence:
> A SaaS for selling digital downloads with Stripe and user accounts

AI suggested configuration:
{
  "projectType": "backend",
  "backendFramework": "express",
  "database": "postgresql",
  "orm": "prisma",
  "auth": "jwt",
  "modules": ["email", "redis"],
  "payments": "stripe"
}

Dependency Philosophy

LaunchStack never pins outdated versions. Dependencies are always installed at their latest stable versions at the time you run the CLI:

# LaunchStack runs this — no version pinning
npm install express prisma @prisma/client jsonwebtoken

Before installation, LaunchStack:

1. Resolves the latest version of each package from the npm registry for display
2. Runs a freshness guard that warns if any package hasn't been updated in over 12 months
3. Checks for and blocks deprecated packages (e.g. request, node-sass)

Template System

Templates are automatically synced from the remote registry at https://github.com/launchstack/templates.

They are cached locally at ~/.launchstack/templates. On each scaffold, LaunchStack:

1. Checks if a local cache exists
2. If online, pulls the latest templates (git pull)
3. If offline, uses the local cache
4. If no cache exists and the registry is unreachable, LaunchStack falls back to the bundled templates included with the CLI.

Templates can be updated independently of CLI releases — no CLI update needed to get the latest template improvements.

# Manually update templates
launchstack templates update

launchstack.json

Projects generated by LaunchStack include a small metadata file used by CLI commands such as launchstack add.

{
  "projectType": "backend",
  "framework": "express",
  "database": "postgresql",
  "orm": "prisma",
  "modules": ["redis", "email"]
}

This file allows LaunchStack to understand the structure of the current project and safely apply additional modules later.

Example Generated Backend Project

my-api/
  src/
    controllers/
    routes/
    services/
    middleware/
    utils/
    config/
  tests/
  .env.example
  Dockerfile
  docker-compose.yml
  launchstack.json
  README.md

The exact structure depends on the selected stack (framework, modules, and Docker configuration).

Environment Variables

After scaffolding, copy .env.example to .env and fill in your values:

cp .env.example .env

Common variables generated based on your stack:

Docker

If Docker is detected, LaunchStack will offer to generate:

• Dockerfile — multi-stage Node.js or Python build
• docker-compose.yml — app + database + Redis (based on your stack)
• .dockerignore

docker-compose up -d

License

MIT