A powerful AI-powered codebase exploration and debugging tool that helps developers understand, analyze, and interact with their repositories through semantic search and intelligent chat assistance.
Features vector-based similarity search using OpenAI embeddings, context-aware AI chat conversations, VAPI-powered voice interactions, and comprehensive repository analytics with AI-generated code detection analysis.

• Hackathon SubmissionDevpost
• Githubgithub.com/AryanBagade/vibe-debugger
• LanguagesTypeScript, JavaScript, SQL
• FrontendNext.js 15.5, React, Tailwind CSS 4.0, Lucide React
• BackendNextAuth.js 5.0, Prisma ORM, Neon DB (PostgreSQL)
• AI & MLOpenAI GPT-4, text-embedding-ada-002, VAPI Voice Assistant
• DatabasePostgreSQL (Neon DB), Qdrant Vector Database
• InfrastructureDocker, GitHub API, Octokit
• Key FeaturesSemantic Code Search, AI Chat Assistant, Voice Integration, Repository Analytics, Smart Indexing

AI-Powered Codebase Exploration Platform

"Making code exploration intelligent and intuitive"

🔍 Core Features

Semantic Code Search: Vector-based similarity search using OpenAI embeddings enables finding code by meaning, not just keywords. Supports 30+ programming languages with precise file location and line number results.

AI-Powered Chat Assistant: Context-aware conversations about your codebase with code explanation, debugging suggestions, architecture insights, and refactoring recommendations - all integrated with repository knowledge.

Voice Assistant Integration: VAPI-powered voice interactions enable hands-free codebase exploration with voice-activated feature introductions and real-time conversation capabilities.

Repository Analytics: AI-generated code detection analysis, commit timeline visualization with contributor insights, repository health metrics, and interactive charts and statistics.

🏗️ Technical Architecture

Built with modern full-stack architecture using Next.js 15.5 with App Router, TypeScript for type safety, and Tailwind CSS 4.0 for styling.

The backend leverages Neon DB (serverless PostgreSQL) for metadata storage, Qdrant vector database for semantic search, and Prisma ORM for database management.

AI capabilities are powered by OpenAI GPT-4 for chat completions, text-embedding-ada-002 for embeddings, and VAPI for voice assistant integration.

🎯 Smart Repository Management

The system handles repositories of varying sizes intelligently:

• Small repos (< 100 files): Full indexing with complete analysis
• Medium repos (100-800 files): Smart filtering with optimized processing
• Large repos (> 800 files): Priority-based selection (500 files max)

Features automated code indexing with progress tracking, GitHub OAuth integration, and optimized handling for both public and private repositories.