BreatheClean.AI

BreatheClean.AI uses OpenAI to answer health questions, forecast air quality, and assess pollen risk based on weather, pollution, and native plant data.

Comment

Inspiration

Poor air quality has become a growing concern across the globe, especially with rising allergies, asthma, and pollution-related illnesses. I wanted to create something that not only informs but protects — something smart, accessible, and science-backed. That’s what led to the birth of BreatheClean.AI

What it does

BreatheClean.AI is a smart environmental health assistant. It:

  • Tracks real-time weather and air pollution (PM2.5)
  • Analyzes native plant species to estimate local pollen risk
  • Gives personalized daily health tips
  • Uses OpenAI's GPT to answer environmental and allergy-related questions in natural language

How we built it

  • Built with Streamlit for UI
  • Used the OpenWeatherMap API for weather and air quality
  • Pulled plant data from the GBIF biodiversity API
  • Integrated OpenAI GPT-3.5 for smart conversational responses
  • Added CSS styling for a modern, mobile-responsive layout

Challenges we ran into

  • Calibrating pollen spread logic using temperature + plant types
  • Optimizing AI prompts to provide medically responsible answers
  • Styling Streamlit components to look professional

Accomplishments that we're proud of

  • Successfully integrated multiple APIs (OpenWeatherMap, GBIF, and OpenAI GPT-3.5) into one cohesive app
  • Built a unique pollen risk estimator using real-time temperature and native plant species data
  • Created a fully functional AI health assistant using OpenAI to answer allergy and air quality questions
  • Developed a clean, interactive Streamlit dashboard with custom CSS for a modern UI
  • Built and deployed the app during Fusion Hacks 2, while balancing multiple other commitments

What we learned

  • Learned how to work with multiple external APIs and handle real-time data integration
  • Gained experience using OpenAI's GPT-3.5 for building natural language features into apps
  • Discovered how to estimate pollen risk by combining plant species data with weather conditions
  • Improved front-end design skills using custom CSS with Streamlit
  • Learned how to scope, structure, and ship a functional MVP in a limited timeframe

What's next for BreatheClean.AI

  • Add voice input so users can speak questions or symptoms instead of typing
  • Launch a mobile-friendly version for easier access on the go
  • Integrate location-based alerts for pollen surges or poor air quality
  • Fine-tune the AI assistant with more health datasets for personalized advice
  • Expand global coverage by supporting multiple languages and more plant regions

Built With

  • gbif-biodiversity-api-(native-plants)
  • html/css
  • openai-gpt-3.5-(ai-assistant)
  • openweathermap-api-(weather-&-air-quality)
  • python
  • streamlit
  • timezonedb-api-(local-time)
Share this project:

Updates