👁️ Vision — AI Visual Assistant (Web Prototype)

Vision is a browser-based AI assistant designed to support accessibility, daily organization, and voice-first interaction, with a focus on assisting visually impaired users. It combines voice recognition, text-to-speech, local data storage, and Google Gemini AI to provide an all-in-one personal assistant experience—entirely in the browser.

This project is currently a front-end prototype intended for learning, experimentation, and future expansion.

---

✨ Features

🚨 Emergency Assistance

• One-tap emergency activation
• Custom emergency contact and message
• Audio confirmation and visual alert feedback

Note: This is a simulation. No real calls or messages are sent.

🎤 Voice Assistant (AI-Powered)

• Speech-to-text via browser APIs
• AI-generated responses using Google Gemini
• Spoken responses with adjustable voice speed
• Context-aware assistance (schedule, health, items, etc.)

📅 Schedule Manager

• Add, view, and delete events
• Date, time, and notes support
• Persistent storage using localStorage

❤️ Health Tracker

• Log health metrics (e.g. blood pressure, medication, heart rate)
• Timestamped entries
• Quick review of recent records

💰 Finance Manager

• Track income and expenses
• Categories and notes
• Automatic balance summary

📍 Item Locator

• Save item locations (e.g. keys, wallet)
• Search items by name
• Spoken location feedback when an item is found

⚙️ Settings

• Emergency contact configuration
• Emergency message customization
• Voice speed control for speech synthesis

---

🧠 Technologies Used

• HTML5 / CSS3 / Vanilla JavaScript

• Web Speech API

  • SpeechRecognition
  • SpeechSynthesis

• Google Gemini API

• Browser Local Storage

• Responsive UI with accessibility considerations

---

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/your-username/vision-ai-assistant.git
cd vision-ai-assistant

2. Add Your Gemini API Key

Open the HTML file and replace:

const GEMINI_API_KEY = 'YOUR_GEMINI_API_KEY_HERE';

with your actual API key from Google AI Studio.

3. Run the App

Simply open the HTML file in a modern browser:

• Chrome (recommended)
• Edge
• Brave

⚠️ Voice recognition requires a Chromium-based browser.

---

⚠️ Limitations & Notes

• This project runs entirely client-side
• No backend or authentication
• Emergency functionality is simulated
• Voice recognition is browser-dependent
• Data is stored locally per browser/device
• Not intended for medical or emergency-critical use

---

🔒 Privacy

• No user data is sent anywhere except:

  • Voice prompts sent to Google Gemini (when enabled)

• All personal data is stored locally in the browser

---

📌 Project Status

Prototype / Proof of Concept

Planned future improvements:

• Backend integration
• Real emergency service hooks
• Camera-based visual assistance
• Offline mode
• User accounts and cloud sync

---

📄 License

This project is released under the MIT License. You are free to use, modify, and distribute it with attribution.

---

🙌 Acknowledgments

• Google Gemini API
• Web Speech API
• Accessibility-first design principles