RepoMind — screenshot 1
RepoMind — screenshot 2
RepoMind — screenshot 3
RepoMind — screenshot 4

RepoMind

"Think inside the repository. Where codebases come alive. One question. Every answer. Zero confusion."

Codex Community Hackathon - Pune

Additional info

How was your experience building with Codex?

Was awesome, just good prompts and it built everything properly

Describe your experience using Loops House as the hackathon platform. What worked well, what challenges (if any) did you face, and what improvements would you like to see?

Very good

Tell us about your overall experience at Codex Community Hackathon Pune.

Very good

What could Codex Community improve to create a better experience for participants?

Everything was good

Team

1 member
  • YU

    Yug Jain

    Owner

Overview

Understanding a new codebase is one of the hardest challenges in software development. Whether you're onboarding at a new job, contributing to open source, or debugging legacy code — existing tools like grep, GitHub search, or Copilot only scratch the surface. RepoMind goes deeper. It clones any GitHub repository, parses it at the function and class level using Tree-sitter, and indexes it using a hybrid of semantic vector embeddings and BM25 keyword search — the same approach used by leading enterprise search systems. Ask any question in natural language and a team of 8 specialized AI agents including a Code Analyst, Security Engineer, Dependency Mapper, and a Principal Reviewer collaborate to deliver a precise, confident answer with exact file:line citations. Different questions route to different expert agents via a LangGraph workflow, so a security question gets a security specialist, not a generalist. Built for production: Redis caching, Celery background jobs, Docker, Railway, Vercel, and GitHub Actions CI RepoMind isn't a prototype. It's the foundation for the future of AI-assisted code understanding.