Inspiration
Aviation weather reports which are called METARS are notoriously hard to read for new pilots or hobbyists. We wanted to make them more accessible, understandable to these people and tose who even have no interest in aviation whil incorporoting Gemini AI to help the end-users further their understanding.
What it does
AeroBrief fetches raw METAR reports from https://aviationweather.gov/ and:
- Translates them into plain English
- Lets users get a quick calculation of wind components
- Summarizes conditions and the AI briefs about the airport.
- Lets users chat with an AI assistant to ask follow-up questions (“Which runway is best?”, “What does BKN120 mean?”) ## How we built it We built this using the gemini studio whch put a nice front-end and back-end which we then adjusted to provide accurate information to aviators. Frontend: React + TypeScript + TailwindCSS
APIs: NOAA/AviationWeather APIs for live METAR/TAF data
AI: Google Gemini integrated for weather decoding, explanations, and chat assistant
Logic: Custom parsers for extracting key fields (wind, visibility, obs time, ceilings) with fallbacks to ensure accuracy
Deployment: Vite + Node.js
Challenges we ran into
The website had to have a good way of interpreting like how visibilty, air pressure etc. is expressed in different parts of the world. We also had issues translating the time of the report in plain english and couple of issues with the UI that were fixed.
Accomplishments that we're proud of
We've built a pretty accurate end-to-end AI weather interpreter in a short amount of time. We solved tricky parsing issues (observation time, runway wind components). All of this with a nice UI with both technical accuracy and visual appeal.
What we learned
How to integrate Gemini AI into a React/Tailwind project effectively The importance of data validation as even in an API the data can be so much different. How to balance aviation accuracy with AI flexibility in prompt design. Working under a time crunch
What's next for AeroBrief
Add TAF (forecast) translation alongside METAR Refine runway recommendation logic with airport runway length, Notices t airmen, and limitations. Deploy a lightweight app so pilots can use it offline/in-flight
Log in or sign up for Devpost to join the conversation.