🤖 Personal AI Assistant — Aria

Voice · Chat · Smart Commands · Powered by Gemini 2.5 Flash

"Talk to it. It listens. It thinks. It acts."

A zero-dependency, single-file personal AI assistant that understands your voice, answers your questions with Gemini AI, checks live weather, sets timers, opens apps, searches the web — and talks back.

🚀 Live Demo  ·  🐛 Report Bug  ·  💡 Request Feature

---

⚡ Features at a Glance

Feature	Description
🎤 Voice Input	Real-time speech recognition (Chrome / Edge)
🔊 Voice Output	Text-to-speech — Aria talks back to you
🧠 Gemini AI	Handles any open-ended question intelligently
🌤️ Live Weather	GPS-based forecast — no extra API key needed
⏰ Smart Timers	Set countdowns with desktop push notifications
📞 Call / SMS	Direct phone & SMS intent links
💬 WhatsApp	Pre-filled message launcher
🐦 Twitter / X	One-command tweet composer
🔍 Web Search	Instant Google search by voice
▶️ YouTube	Search & play videos hands-free
🗺️ Google Maps	Navigate anywhere by voice
📱 PWA Ready	Installable on Android / iOS home screen
🔒 Fully Private	No server · No storage · Runs in your browser

---

🗂️ Project Structure

Nova/
├── index.html       ← Entire app (single self-contained file)
├── manifest.json    ← PWA manifest for mobile install
├── icon-192.png     ← PWA icon
└── README.md        ← You are here

No build tools. No npm. No frameworks. One file is all it takes.

---

🚀 Getting Started

Step 1 — Get the Code

git clone https://github.com/Abd-Abdullah83/Nova.git
cd Nova

Or just download index.html and open it — it runs offline.

Step 2 — Get a Free Gemini API Key

1. Go to aistudio.google.com/app/apikey
2. Sign in with your Google account
3. Click "Create API Key" and copy your key

Step 3 — Paste Your Key

Open index.html and find this line near the top of the <script>:

const API_KEY = "YOUR_KEY_HERE";

Replace YOUR_KEY_HERE with your actual key.

Step 4 — Open & Use

# Simply open in your browser:
open index.html

# Or serve locally for full PWA features:
npx serve .

⚠️ Chrome or Edge only — Firefox does not support the Web Speech API.

---

🗣️ Command Reference

🕐 Time & Date

Say	Result
"What's the time?"	Current time in your locale
"What's today's date?"	Full localized date string

🌤️ Weather

Say	Result
"What's the weather?"	Live weather from your GPS location
"What's the temperature?"	Temperature, humidity, wind speed

⏰ Timers

Say	Result
"Set a 5 minute timer"	Countdown + desktop notification
"Set a 30 second timer"	Works with seconds, minutes, hours
"Set an alarm"	Defaults to 5 minutes

🔍 Search & Navigation

Say	Result
"Search for FAST University"	Google search
"Open YouTube"	Launches YouTube
"Play lo-fi music on YouTube"	YouTube search query
"Navigate to Lahore"	Google Maps directions
"Open github.com"	Opens any website

💬 Messaging & Social

Say	Result
"Send WhatsApp message hello"	WhatsApp with pre-filled text
"Post on Twitter great day"	Twitter compose
"Open Instagram"	Opens Instagram

📞 Communication

Say	Result
"Call +923001234567"	Phone dialer
"SMS +923001234567 saying hello"	SMS with number and body

🧠 AI Chat — Gemini Fallback

Say	Result
"Explain machine learning"	Conversational AI answer
"Tell me a joke"	Gemini responds with humour
"Write a short poem"	Creative AI response

---

🛠️ Customisation

Rename the Assistant

systemInstruction: {
  parts: [{ text: "You are a helpful assistant named Nova. Be friendly." }]
}

Add a New Voice Command

if (q.includes("calculator")) {
  respond("Opening calculator!");
  window.open("https://www.google.com/search?q=calculator");
  return;
}

Change AI Personality

"You are a sarcastic but helpful assistant. Keep it witty."
"You are a professional data science tutor."
"You only respond in Urdu."

Switch Language

recognition.lang = "ur-PK";   // Urdu
recognition.lang = "ar-SA";   // Arabic
recognition.lang = "fr-FR";   // French

---

🔒 Privacy & Security

Concern	Reality
Is data stored?	❌ No — nothing is ever saved anywhere
Is there a server?	❌ No — runs entirely inside your browser
API key exposure	⚠️ Key is client-side — for personal use only
Weather & location	GPS coords sent only to Open-Meteo's public API
Voice data	Processed locally by your browser's speech engine

---

🐛 Troubleshooting

Problem	Solution
Mic button does nothing	Switch to Chrome or Edge
"Mic error — check permissions"	Allow microphone in browser site settings
AI gives no response	Verify your API_KEY is valid and has quota
Weather not loading	Allow location access when browser prompts
Links / popups blocked	Allow popups for this page in browser settings
Timer notification silent	Allow notifications when prompted

---

📊 Tech Stack

Layer	Technology
UI	Vanilla HTML5 + CSS3 (dark theme, purple accent)
Logic	Vanilla JavaScript ES2020+
AI Brain	Google Gemini 2.5 Flash (REST API)
Voice In	Web Speech API — SpeechRecognition
Voice Out	Web Speech API — SpeechSynthesis
Weather	Open-Meteo (free, no key) + Geolocation API
Alerts	Notifications API + setTimeout
Installs as App	PWA via manifest.json

---

📄 License

MIT License — Free to use, modify, and distribute. Give credit where it's due. ✌️

---

── Built by ──

Abdullah Tahir

BS Data Science · FAST NUCES Lahore · 2025–2029

"Build a company of lasting impact." 🚀

⭐ Star this repo if it helped you — it means a lot!

Made with 💜 and a lot of late-night coding sessions