🌍🎙️ GeoAssistant: Location-Aware Voice-Enabled Tourist Assistant

GeoAssistant is a Progressive Web App (PWA) designed to function as a location-aware, voice-enabled assistant for tourist spots. It detects the user's GPS location, identifies nearby landmarks, and responds with voice-guided information using a Zephyr-7B Alpha-powered chatbot hosted via FastAPI.

---

📝 Project Overview

GeoAssistant is a proof-of-concept smart assistant that can:

• Automatically detect when a user is near a tourist destination
• Provide spoken prompts and suggestions
• Accept voice or text queries
• Return context-aware responses using a powerful language model
• Function as a PWA, accessible via any modern mobile browser

🔍 Note: This repository contains only the essential files to demonstrate the core functionality. Supporting assets, helper scripts, and extended configurations will not be uploaded to GitHub.

---

📸 Screenshot of the Web App

Here’s a screenshot of the Location-Aware Voice-Enabled Assistant web app in action:

Note: The coordinates for Gateway of India in the image are not the real ones. To ensure the GeoAssistant works correctly, I saved my live coordinates as "Gateway of India" in the locations.json file (since my house is not a tourist spot). This way, when it fetched my real-time location, it displayed "Gateway of India" with the corresponding description. You can add any other location instead of Gateway of India for testing or deployment purposes. 🗺️📍

---

📂 File Descriptions

### 🔧 Backend Files

backend/server.py

A FastAPI server that:

• Receives user queries (optionally with GPS location)
• Passes the prompt to the Zephyr-7B Alpha language model
• Returns the generated response to the client
• Powers the chatbot logic behind the voice assistant

backend/requirements_backend.txt

Lists essential Python dependencies for the backend:

• FastAPI for API handling
• PyTorch (GPU-supported) for Zephyr model
• Speech Recognition and Text-to-Speech libraries
• Ensures smooth functioning of the voice-enabled backend

---

🌍 Frontend Files

frontend/locations.json

A JSON database of landmarks:

• Defines coordinates and radii for tourist spots
• Used to trigger the chatbot or voice guide when user is nearby

frontend/gps.js

Fetches the user’s current GPS location:

• Uses the browser’s built-in geolocation API
• Integrates with location trigger logic

frontend/geo_trigger.js

Performs proximity detection:

• Compares user GPS location to locations.json
• Uses Haversine formula to calculate distance
• Triggers chatbot/audio if user is within specified radius

frontend/App.jsx

The core React frontend of the PWA:

• Detects and tracks user’s location
• Checks for proximity to landmarks
• Enables voice or text queries
• Speaks model responses using Text-to-Speech
• Shows the user on an interactive Leaflet map
• Allows toggling of location & voice features

---

📦 Project-wide Requirements

requirements_all.txt

A comprehensive package list for the full stack:

• Frontend: React + location/voice support
• Backend: FastAPI, PyTorch, Transformers
• Optional ML tools: Jupyter, TensorFlow/Keras, etc.
• Ensures compatibility across all modules

---

🚀 Features at a Glance

• 📍 Live GPS tracking to detect nearby tourist spots
• 🔄 Location triggers for auto-activation of the assistant
• 🗣️ Voice input/output for a hands-free experience
• 💬 Chatbot integration using Zephyr-7B Alpha
• 🌐 Works as a PWA, installable on any smartphone
• 🗺️ Map view using Leaflet to visualize user position

---