⚡ AutoBlog: Full-Stack Automated Content Platform

This repository contains the solution for the Full-Stack Technical Challenge, focusing on building and deploying a fully automated blog application with AI-powered content generation.

The application features a modern frontend for viewing articles, a Node.js backend with AI integration for content creation, and a production-ready CI/CD pipeline leveraging AWS ECR, CodeBuild, and EC2.

📐 Architecture Overview

The system is a fully containerized, microservice-like architecture composed of three main services orchestrated by Docker Compose:

1. Frontend – React/Vite user interface served on port 80
2. Backend – Node.js/Express API with database and AI scheduler on port 3003
3. Proxy – Nginx reverse proxy routing traffic to appropriate services

Traffic Flow: User (Port 8080) → Nginx Proxy → Routes to Backend (/api) or Frontend (/)

---

🛠 Tech Stack

Frontend

• React with TypeScript and Vite
• Tailwind CSS for styling
• React Router for client-side navigation
• React Markdown with remark-gfm and rehype-raw for secure Markdown rendering

Backend & Data

• Node.js with Express.js
• SQLite for persistent article storage (mounted via Docker Volume)
• AI/LLM: Gemini 2.5 Flash via LangChain and @langchain/google-genai

DevOps & Infrastructure

• Docker with multi-stage builds for optimized images
• Docker Compose for local and EC2 orchestration
• Nginx reverse proxy
• AWS ECR, CodeBuild, and EC2 for CI/CD and deployment

---

💡 Key Technical Decisions

1. Robust AI Content Generation

The content generation logic in backend/services/aiAgent.js implements structured output for reliability:

• Uses a Zod schema (ArticleOutputSchema) to strictly define article fields: title, content, summary, and author
• The @langchain/google-genai module is configured with .withStructuredOutput(ArticleOutputSchema) to enforce valid JSON responses from the LLM, preventing malformed database entries
• An external tool fetches live trending topics before generating articles

2. Automated Scheduling

• The backend uses node-cron to automatically generate and save a new article daily at 1:00 AM UTC
• The generation function also runs on server startup, ensuring the application has content available immediately

3. CI/CD Pipeline (AWS)

CodeBuild executes the following steps via infra/buildspec.yml:

1. Authenticate with AWS ECR
2. Build frontend and backend images using multi-stage Dockerfiles
3. Tag images with both latest and a unique commit hash
4. Push tagged images to their respective ECR repositories

EC2 Deployment: The infra/scripts/deploy.sh script runs on the EC2 instance to:

1. Authenticate with ECR
2. Pull the latest images
3. Deploy the entire stack using docker-compose up -d --force-recreate for zero-downtime updates

---

🚀 Local Development Setup

To run the application locally, you need Docker and Docker Compose installed.

Prerequisites

Create a .env file in the project root with the following variables:

AWS_REGION=<Your_AWS_Region>
AWS_ACCOUNT_ID=<Your_AWS_Account_ID>

FRONTEND_IMAGE_NAME=autoblog-frontend
BACKEND_IMAGE_NAME=autoblog-backend

GEMINI_API_KEY="YOUR_GEMINI_API_KEY"
HASDATA_API_URL="https://api.hasdata.com/scrape/google/news"
HASDATA_API_KEY="YOUR_HASDATA_API_KEY"

Running the Application

1. Build and run the application:

   docker-compose up --build

   This command builds the images, creates the Docker network, and starts the frontend, backend, and proxy containers.

2. Access the application: Open your browser and navigate to http://localhost:8080

Database Persistence

The SQLite database (blog.sqlite) is stored in a named Docker volume (blog-data) defined in docker-compose.yaml. This ensures articles persist across container restarts and updates.

---

📝 Notes

• Ensure all required API keys are set before running the application
• The database volume is automatically created and managed by Docker Compose
• For production deployment, update the environment variables in your EC2 instance or AWS Systems Manager Parameter Store

📄 License

This project is licensed under the MIT License.