Skip to main content
Version: Capstone Spring 2025 (Archived)

Component Overview

The AI-assisted coding assistant project consists of several key components and technologies, including a VS Code extension for code suggestions, a Go backend for processing and logging user interactions, Gemini as the AI model, Supabase for authentication and database storage, and a React.js dashboard for users to track their progress.

Figure 1. Component Diagram of the CLOVER application.

Main Application​

🛠 VS Code Extension (TypeScript, Node.js)​

The core of the system is a VS Code extension that provides code suggestions while occasionally introducing small logic errors to test user attentiveness.

🔹 Key Features:​

  • AI-Powered Code Suggestions – Uses Gemini to generate helpful code recommendations.
  • Bug Injection – Sometimes modifies code (e.g., changing add(a, b) to a - b) to check if users notice mistakes.
  • User Response Logging – Tracks how users interact with suggestions (accept, modify, or reject).
  • Adjustable Difficulty – Adapts suggestions based on past responses.

Backend Services​

🌐 Gin (Golang)​

The backend handles all API requests between the VS Code extension, AI model, and database.

🔹 Key Features:​

  • Processes Code Requests – Sends user code to the AI model and returns suggestions.
  • Tracks User Behavior – Logs whether users accept, modify, or reject suggestions.

AI Model​

🤖 Gemini (AI Code Generator)​

The AI model generates code suggestions and sometimes introduces small mistakes to test users.

🔹 Key Features:​

  • Context-Aware Suggestions – Provides relevant recommendations based on the user’s code.
  • Intelligent Mistakes – Occasionally tweaks suggestions with logical errors to test user focus.
  • Adaptive Learning – Adjusts suggestions based on user behavior.

Database & Authentication​

🔐 Supabase (PostgreSQL, Auth)​

Supabase handles user login and stores all interaction data.

🔹 Key Features:​

  • User Authentication – Manages logins and keeps track of individual progress.
  • Logs User Activity – Records which suggestions were accepted, modified, or rejected.
  • Real-Time Sync – Updates and injects user progress instantly to the AI model and on the dashboard.

User Dashboard​

📊 React, TailwindCSS, and ShadCN​

The dashboard allows users to view their progress and track their learning journey.

🔹 Key Features:​

  • Progress Tracking – Displays stats on correct vs. incorrect responses.
  • Insightful Feedback – Helps users understand where they need improvement.
  • Encourages Learning – Provides interactive insights to refine coding skills.