Random5-01

asdf

Comment

Inspiration

Creating an intelligent assistant that helps users discover places, visualize them through relevant images, and make weather-informed decisions.

What it does

It's an AI-powered travel and location assistant that:

  • Understands natural language queries about places (restaurants, cafes, tourist attractions, etc.)
  • Extracts user preferences automatically (ratings, distance, accessibility)
  • Searches across multiple location APIs to find the best matches
  • Scrapes websites and uses AI to filter the most relevant images based on context
  • Provides weather forecasts with natural language processing

How we built it

  • Backend: FastAPI for high-performance REST API -Frontend: Streamlit, open-source Python framework to deliver interactive apps
  • Web Scraping: BeautifulSoup for extracting images from websites
  • Location Services: Tested and integrated different place APIs (Google Places, Overpass, Yelp, etc.)
  • Weather Data: Open-Meteo API for hourly forecasts with geocoding via OpenStreetMap

Challenges we ran into

  • Integrating multiple place APIs with different data formats and rate limits
  • Filtering out irrelevant images (logos, icons, SVGs) from scraped websites

Accomplishments that we're proud of

-We didn't give up (almost :D)

What we learned

  • GPT prompt engineering for consistent JSON output and context-based selection
  • Working with geospatial data and coordinate systems
  • Web scraping techniques
  • Integrating tools to enable LLMs to perform actions via different APIs
  • Using agent loop control to manage multi-step reasoning and repeated tool invocations

What's next for Random5-01

  • Add more filtering options (opening hours, specific amenities)
  • Add user preferences storage and personalized recommendations
Share this project:

Updates