My Projects

Overview Travel Helper is a free, open-source web application that helps users discover places to eat, rest, workout, or explore when traveling in any city worldwide. It combines map-based search with AI-generated descriptions, curated photos, and enriched venue metadata. Key Features 1.Category-based venue discovery: Find restaurants, cafés, hotels, parks, gyms, museums, etc. 2.Interactive map & geolocation support: View results on a map; use “Near Me” to find nearby spots. 3.AI-generated descriptions: Uses a local LLM (via Ollama) to generate context-aware blurbs about each venue. 4.Free photos & details: Pulls images from Wikimedia Commons and metadata (opening hours, contact, accessibility) from OpenStreetMap. 5.No API keys or usage limits: Designed to run purely on free, open data sources; no credit card or rate-limited APIs required. Tech Stack & Architecture 1.Next.js (App Router) + TypeScript for front-end and server-side rendering 2.Tailwind CSS for styling 3.Leaflet.js for map visualization 4.OpenStreetMap (Nominatim + Overpass API) for venue search and geocoding 5.Wikimedia Commons API for venue photos 6.Ollama / local LLM models (llama3.2, llama2, mistral, etc.) for descriptive text generation 7.Modular structure: separate layers for APIs, components, and data integrations Why It Matters 1.Avoids dependency on paid mapping/POI services (e.g. Foursquare, Google Places) 2.Remains fully functional even if AI is offline (venues still show) 3.Open and extensible — others can contribute or adapt 4.Ideal proof of concept for combining geospatial search + AI in a lightweight, user-friendly app Usage Flow 1.Enter a destination city or click “Near Me” 2.Pick a category (e.g. Eat, Rest, Workout, Explore) 3.Browse results in list or map mode 4.Click any venue to view photos, AI description, contact info, hours, accessibility, etc.