Repoxry - Understand. Analyze. Navigate.

Vaibhav Sengar — Thu, 15 Oct 2026 15:58:50 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

Repo X-Ray is a local-first AI developer tool that helps developers understand unfamiliar codebases.

Instead of manually digging through dozens or hundreds of files, Repo X-Ray analyzes a repository and provides:

Interactive architecture and dependency visualization
AI-powered codebase explanations
File and module analysis
Dependency impact analysis
AI-assisted code health checks
Project architecture summaries
README/documentation generation
Natural-language questions about the codebase

The target users are developers, freelancers, open-source maintainers, and small development teams who regularly need to understand existing projects.

The idea is simple:

Drop in a codebase. See inside it.

Repo X-Ray could eventually be offered as a paid developer tool for individuals and teams who want faster codebase onboarding and analysis without sending their private source code to a third-party AI service.

## Demo

Live Demo: https://repoxry.onslate.in/

The demo shows Repo X-Ray analyzing a real project, generating its architecture map, answering questions about the codebase, and performing impact analysis.

One of the key parts of the demo is running the AI locally through Ollama rather than relying on a cloud AI API.

## Code

GitHub: https://github.com/vaiibhavsengar/REPO-XRAY

The project is built with a local-first architecture and is intended to be open source.

## How I Built It

Repo X-Ray is built around local inference using Ollama and open-weight AI models.

The application uses:

HTML, CSS and Vanilla JavaScript for the frontend
Node.js + Express for the local backend
Ollama for local AI inference
Cytoscape.js for the interactive architecture/dependency graph
Local repository scanning and static analysis for understanding project structure

The basic architecture is:

Local Repository
       ↓
Repo Scanner
       ↓
Dependency / Structure Analysis
       ↓
Relevant Code Context
       ↓
Ollama
       ↓
Open-Weight AI Model
       ↓
Repo X-Ray
       ↓
Architecture / Explanations / Insights

Repo X-Ray does not blindly send an entire repository to the model. The application first extracts useful project context such as files, directories, dependencies, imports and relevant code, then provides the model with the information needed for a specific analysis.

Users can connect their own local Ollama installation and choose an installed model. The application does not require my personal API key or a proprietary cloud AI provider.

## Why Does Open Innovation Matter?

For a developer tool that analyzes source code, privacy matters.

A codebase can contain proprietary business logic, internal architecture, private configuration and other information that a developer may not want to upload to a third-party AI service.

With Repo X-Ray, the AI can run locally:

Developer's Code
      ↓
Local Repo X-Ray
      ↓
Local Ollama
      ↓
Open-Weight Model

The source code does not need to be sent to a cloud AI provider.

Open AI also gives Repo X-Ray another important advantage: model freedom.

Instead of building the product around one proprietary API, users can choose the model that works best for their machine and use case. The AI layer can evolve independently from the rest of the application.

There is also no per-request cloud AI cost when inference is performed locally.

For this project, open innovation isn't just a way to make the AI cheaper. It changes the privacy, deployment and ownership model of the entire product.

## My First User

Yes! My first user was my developer friend, and he is completely shocked after using my saas, yup I charged him around 2000 INR ($20 approx.), He said it was completely worth it.

## Prize Categories

Ollama (Gemma 4)

🌿 Off-Line — AI that helps you leave the screen

Vaibhav Sengar — Wed, 07 Oct 2026 05:00:00 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

🌿 Off-Line — AI that helps you leave the screen

What I Built

Off-Line is a simple outdoor mission experience built around one idea:
AI should not always keep us on our screens. Sometimes it should help us leave them.

The user chooses how much time they have — 15, 30, or 60 minutes — and an activity such as Explore, Nature, Photography, or Surprise Me.
Off-Line then gives them a small real-world mission to complete outside.

For example:
The Unnoticed Route
Take a route you don't normally take. Find three things you've never noticed before.
Once the mission starts, the interface becomes minimal. The goal is to put the phone away and actually complete the mission.
After returning, the user can mark the mission complete, write a short reflection, and earn points.
Completed missions and points are stored locally using localStorage.
The idea isn't to build another app that keeps people scrolling.
The goal is to give people a reason to close the app.

Demo
Live Demo:
[https://offlineformlh.netlify.app/]
The current version runs in Demo Mode, so it doesn't require Ollama, API keys, or any external AI service to experience the complete product.

Code

GitHub Repository:
[https://github.com/vaiibhavsengar/Off-Line]
The project is intentionally built with a simple and understandable stack so that other developers can easily explore and extend it.

How I Built It
Off-Line is currently built using:

HTML
CSS
Vanilla JavaScript
Node.js
Express
LocalStorage The current version uses a local Demo Mode mission generator instead of an external AI API. I intentionally separated the mission-generation logic from the rest of the application so that a personal/local AI layer can be integrated later without rebuilding the entire product. The planned architecture is:

*User
↓
Off-Line UI
↓
Mission Generator
↓
Personal / Local AI
↓
Outdoor Mission
*
The future AI layer can use an open-weight model running locally through Ollama.
For this submission, I am keeping that integration as a future extension rather than pretending the current Demo Mode is AI-powered.

Why Does Open Innovation Matter?
Open innovation matters because it gives developers the freedom to experiment beyond a single closed AI provider.
For Off-Line, the future goal is to allow the mission-generation intelligence to run locally using an open-weight model.
That could make it possible to:

Experiment with different models
Run AI locally
Modify the AI behavior
Avoid depending on a single closed API
Build a more personal AI experience
Keep the core experience open and extensible

More importantly, the project explores an interesting direction for AI:
What if AI wasn't designed to keep us online?
Instead of generating more content for us to consume, AI could generate a reason for us to stop consuming content and go experience something ourselves.

Prize Categories
Overall Hactoberfest Open Source AI Challenge Week 1: Touch Grass
Best Use of Gemma

🌱 Final Thought by me :

Technology doesn't always need to demand more of our attention.
Sometimes the best thing an app can do...
is tell you to put the phone down and go outside.

Project ScreenShots :

DEV Community: Vaibhav Sengar

Repoxry - Understand. Analyze. Navigate.

🌿 Off-Line — AI that helps you leave the screen

Code