🔧 AI-Powered Personal Assistant

An interactive, voice-controlled personal assistant built with a lightweight frontend and a Python-based backend. It leverages cutting-edge LLM technologies and tool-calling capabilities for real-time tasks like Wikipedia summarization, image fetching, task management, and more.

📚 Learn More

• 📝 Technical Deep Dive on Medium - Learn about the optimization techniques and architecture decisions

📸 Image Examples

Below are examples of visual responses generated by the assistant:

Login Page

Weather UI

Song Player

To-Do Manager

News Feed

🚀 Features

• 🌐 Real-Time Wikipedia Summaries
  Retrieves and summarizes Wikipedia content dynamically using FastAPI endpoints.

• 🖼️ Visual Context with Images
  Automatically fetches and displays relevant images alongside answers.

• 🎵 Integrated Music & Task Assistant
  Built-in music controls and Google Tasks API integration for reminders and task management.

• 🌤️ Live Weather and News
  Displays up-to-date weather conditions and trending headlines.

• 🔍 Improved Search Accuracy
  Uses Google Custom Search API to access Wikipedia data with 70% higher factual accuracy.

• ⚙️ FastAPI Backend + Vanilla JS Frontend
  Responsive UI powered by HTML/CSS/JS and a scalable Python backend with FastAPI.

• 🧠 Optimized LLM Performance
  8-bit quantized FineTuned Qwen 2.5B reduces GPU usage by 50%.

• 🧩 Context-Aware Tool-Calling
  Dynamically routes queries to appropriate tools (e.g., wiki, weather, tasks) and handles missing-info prompts intelligently.

• 🧱 Scalable Architecture
  Backend built with Supabase + FastAPI.

---

🛠️ Tech Stack

• Frontend: HTML, CSS, JavaScript (Vanilla)
• Backend: FastAPI (Python)
• LLM: Qwen 2.5B (8-bit quantized)
• Database: Supabase
• APIs: Google Tasks API, Google Custom Search API, OpenWeather API

---