VeridianAI: Personalized AI-Powered Newsletters

Personalized AI Newsletters Tailored for You

---

Overview

VeridianAI is an end-to-end platform that delivers personalized, AI-generated news summaries to users based on their selected topics. The system leverages state-of-the-art NLP models and a modern web stack to automate news aggregation, summarization, and delivery, providing a seamless and engaging user experience.

If you want to see the deployed app, go to VeridianAI. Sign up now and enjoy free newsletters every morning at 8am EST and catch up on your interested topics in a quick bite sized read!

---

Table of Contents

• Features & MVP
• Tech Stack
• Architecture
• Screenshots & Suggested Images
• Getting Started
• Folder Structure
• License

---

Features & MVP

Minimum Viable Product (MVP):

• User Authentication: Secure sign-up, login, and session management (Firebase Auth).
• Topic Selection: Users choose their preferred news topics (e.g., tech, science, sports).
• Automated News Fetching: Daily aggregation of news articles from external APIs.
• AI Summarization: Use of Hugging Face Transformers to generate concise, high-quality summaries.
• Duplicate Detection: Semantic similarity checks to avoid redundant content.
• Personalized Newsletters: Each user receives a custom newsletter with summaries for their chosen topics.
• Newsletter Delivery: Newsletters are stored in Firestore and can be sent via email (Brevo integration).
• Modern UI: Responsive, animated React/Next.js frontend with light/dark mode.

---

Tech Stack

Frontend

• Framework: Next.js (React 19)
• Styling: Tailwind CSS, Framer Motion (animations)
• State/Auth: React Context, Firebase Auth
• Deployment: Vercel

Backend

• Language: Python 3
• AI/NLP: Hugging Face Transformers (facebook/bart-large-cnn), Sentence Transformers (all-MiniLM-L6-v2)
• News Parsing: newspaper3k and thenewsapi
• Database: Google Firestore (via firebase-admin)
• Environment: python-dotenv
• Email Delivery: Brevo (Sendinblue) API

---

Architecture

flowchart TD
    A["User"] -->|"Web UI"| B("Next.js Frontend")
    B -->|"Auth, Preferences"| C["Firebase Auth/Firestore"]
    B -->|"API Calls"| D["Backend Python Pipeline"]
    D -->|"Fetch News"| E["News API"]
    D -->|"Summarize"| F["Hugging Face Transformers"]
    D -->|"Store Summaries"| C
    D -->|"Send Email"| G["Brevo API"]
    C -->|"Personalized Data"| B

Loading

---

Screenshots

Login Screen

Settings Page

Newsletter Early Examples. (Subject to change with each release)

Currently working on implementing the images into the newsletter as well as updating the structure of the newsletter to fit the theme of the website

---

Getting Started

Backend

1. cd backend
2. Install dependencies:

   pip install -r requirements.txt

3. Set up your .env with Firebase and API keys.
4. Run the pipeline:

   python ai_pipeline.py

Frontend

1. cd frontend
2. Install dependencies:

   npm install

3. Start the dev server:

   npm run dev

4. Visit http://localhost:3000

---

Folder Structure

AI Newsletter Project/
  backend/         # Python AI pipeline, summarization, Firestore integration
  frontend/        # Next.js React app, user auth, topic selection, UI

---

License

MIT