TECH STACK:

THE WEBSITE:

THE GOAL:

Looking for a laptop but can't pin down what brand or model fits you best? Stratus was built to help users receive quality recommendations based on their personal needs, without digging through endless reviews or spec lists.

CORE FUNCTIONALITIES:

- RAG-based AI Recommendations: Stratus uses a Sentence Transformer model to provide personalized laptop suggestions using semantic search.
- Fullstack Integration: Next.js for the frontend, Flask + Redis for backend processing and stateful sessions.
- Model-backed Inference: Embeddings and product data are pre-loaded for instant response.

AI STORYTELLING:

Stratus uses the "all-MiniLM-L6-v2" transformer model to encode both user queries and laptop data. It calculates cosine similarity between query and product embeddings to return the top 3 matches. This approach mimics personalized search while keeping inference fast and efficient.

QUICK START:

🔧 PREREQUISITES:

- Python 3.11+ - Node.js 18+ - Redis (cloud or local)

1. CLONE:

git clone https://github.com/YaredPena/stratus.git
cd stratus

2. VENV:

cd server
python -m venv venv
source venv/bin/activate | (Windows: .\venv\Scripts\activate)

3. DEPENDENCIES:

pip install -r requirements.txt

4. ENV CONFIG:

REDIS_URL= your_redis_url

SECRET_KEY= your_flask_secret_key

5. DATABASE SETUP:

Ensure the following files are present in the data/ folder:

-laptops.pkl

-laptop_embeddings.npy

6. RUN THE APPLICATION:

gunicorn app:app --bind 0.0.0.0:5000

FOR FRONTEND

cd ../client
npm install
echo "NEXT_PUBLIC_API_URL=http://localhost:5000" > .env.local
npm run dev