Templates

ForgeKit ships with six production-ready templates. Each template is fully wired with the right stack, Docker support, CI/CD, tests, and environment configuration from the first scaffold. Nothing is a placeholder; every file runs on day one.

web-appTypeScript

Full-stack React + Express application with Vite. One repo, one deploy.

View docs →

api-servicePython

Production REST API with FastAPI, PostgreSQL, SQLAlchemy 2.0, and Alembic migrations.

View docs →

ml-pipelinePython

Reproducible ML workflow with MLflow tracking, Jupyter Lab, and a structured source layout.

View docs →

next-appTypeScript

Next.js App Router with Tailwind CSS, TypeScript, and Docker support out of the box.

View docs →

go-apiGo

High-performance REST API built with Go and the Gin framework, wired with PostgreSQL.

View docs →

serverlessAWS

TypeScript Lambda functions managed with the Serverless Framework. Deploy without managing servers.

View docs →

Available Templates

Template	Stack	Language	Use Case
web-app	Node.js + React + TypeScript + Express + Vite	TypeScript	Full-stack web application
api-service	Python + FastAPI + PostgreSQL + SQLAlchemy 2.0	Python	REST API service
ml-pipeline	Python + Jupyter + MLflow + scikit-learn	Python	Machine learning workflow
next-app	Next.js + TypeScript + Tailwind CSS	TypeScript	Full-stack Next.js application
go-api	Go + Gin + PostgreSQL + Docker	Go	High-performance REST API
serverless	TypeScript + AWS Lambda + Serverless Framework	TypeScript	Serverless functions

---

Which template should I choose?

Use this decision guide to pick the right starting point.

I want to build a web app with a UI

• I need React and a Node.js API in one repo → use web-app
• I want Next.js with Tailwind CSS and server-side rendering → use next-app

I want to build a backend API

• I'm a Python developer and need a REST API with a database → use api-service
• I want high performance and am comfortable with Go → use go-api
• I want functions that run in the cloud without managing servers → use serverless

I want to do machine learning

• I need reproducible experiments, Jupyter notebooks, and MLflow tracking → use ml-pipeline

I'm not sure

Run the interactive wizard and it will prompt you:

npx forgekit-cli new

---

What every template includes

• Dockerfile for building a production container image
• docker-compose.yml for local development
• .env.example with documented environment variables
• .gitignore configured for the stack
• README.md with project-specific setup instructions
• A health check endpoint
• Tests for core functionality
• GitHub Actions CI workflow

---

Template details

web-app

Full-stack web application with a React 18 frontend (Vite + TypeScript) and an Express 4 backend.

Best for: Projects that need both a UI and an API in one repository.

Quick start:

npx forgekit-cli new my-app --template web-app
cd my-app
npm run dev

Frontend runs at http://localhost:5173. Backend runs at http://localhost:3000.

Full web-app documentation →

---

api-service

Production-ready REST API with FastAPI, SQLAlchemy 2.0 ORM, Alembic migrations, Pydantic v2 validation, and a PostgreSQL database.

Best for: Python developers building standalone REST APIs consumed by other services or frontends.

Quick start:

npx forgekit-cli new my-api --template api-service
cd my-api
docker-compose up --build

API runs at http://localhost:8000. Auto-generated docs at http://localhost:8000/docs.

Full api-service documentation →

---

ml-pipeline

Reproducible machine learning workflow with a structured source layout (data loading, feature engineering, training, evaluation), MLflow experiment tracking, Jupyter Lab, and a Makefile for common tasks.

Best for: Data scientists and ML engineers who need a structured, reproducible project from day one.

Quick start:

npx forgekit-cli new my-ml-pipeline --template ml-pipeline
cd my-ml-pipeline
pip install -r requirements.txt
make train

MLflow UI runs at http://localhost:5000 when you start it with mlflow ui.

Full ml-pipeline documentation →

---

next-app

Full-stack Next.js application with the App Router, TypeScript, Tailwind CSS, and a built-in health API route. Includes Docker support for production deployments.

Best for: Teams who prefer the Next.js full-stack model with server components and Tailwind for styling.

Quick start:

npx forgekit-cli new my-next-app --template next-app
cd my-next-app
npm run dev

App runs at http://localhost:3000.

Full next-app documentation →

---

go-api

High-performance REST API built with Go and the Gin framework. Includes PostgreSQL integration, a Makefile for common Go tasks, and Docker support.

Best for: Engineers who want Go's performance and concurrency for a backend API.

Prerequisites: Go 1.21 or higher must be installed (go version).

Quick start:

npx forgekit-cli new my-go-api --template go-api
cd my-go-api
docker-compose up --build

API runs at http://localhost:8080.

Full go-api documentation →

---

serverless

AWS Lambda functions written in TypeScript, managed with the Serverless Framework. Includes a health check handler, a sample function, and tests.

Best for: Teams deploying event-driven workloads or APIs to AWS Lambda without managing infrastructure.

Prerequisites: An AWS account and AWS credentials configured locally for deployment.

Quick start:

npx forgekit-cli new my-functions --template serverless
cd my-functions
npm run dev   # Local invocation via Serverless offline

---

Full serverless documentation →

---

Template Pages

• Web App
• API Service
• ML Pipeline
• Next App
• Go API
• Serverless