Airomas

Inspiration

I’ve always been fascinated by fragrances and how scents can evoke memories, moods, and personality. I wanted to create a tool that helps users discover perfumes and colognes tailored to their unique traits, style, and lifestyle. Everybody should smell great.

What it does

Airomas is an AI-powered fragrance recommender and chat assistant. Users can input personal details—age, gender, mood, activities, and style—and get personalized scent suggestions. It also lets users chat with an AI about fragrances, trends, and general advice.

How I built it

I built Airomas using Flask for the web backend and OpenRouter’s Google Gemini 2.5 AI for fragrance recommendations and chatbot responses. The front end uses HTML, CSS, and JavaScript to create a clean, responsive chat interface with message bubbles and markdown support. Data is dynamically formatted into JSON for consistent display.

Challenges I ran into

Linking the project to a .tech domain. Working with GIT Ensuring the links generated worked, a problem to be addressed. Uploading my API keys by accident...

Accomplishments that I'm proud of

Successfully integrated AI-powered recommendations with dynamic front-end display. Implemented links for each fragrance so users can explore products immediately. Built a fully functional chat interface with markdown support, proper message bubbles, and responsive styling.

What I learned

How to prompt AI effectively to get structured JSON output. How to link a project to a domain through Vercel. Not leaking my API keys...

What's next for Airomas

Ensure the links work and direct users to the appropriate website. Expand e-commerce integration for users to purchase fragrances through my site. Improve the chatbot with more personalized and fine-tuned responses.

DISCORD: chee280