This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
What I Built
OpenPath is a lazy person's way into open source.
Making a first contribution means a lot of dreaded homework: scrolling through hundreds of issues, guessing which ones fit your skills, decoding jargon-heavy descriptions, and figuring out what to do first. Most beginners give up before writing a single line of code.
OpenPath does that homework for you:
- Finds the issues: paste any public GitHub repo URL and it pulls in the real issues.
- Picks for you: tell it your skills and experience level, and it recommends issues that are potentially a good fit, using issue details and labels.
- Explains them: local AI breaks down what an issue is about, which skills matter, and why it could be a good starting point.
- Plans it for you: you get a step-by-step roadmap: understand the issue, read the contribution guidelines, inspect the code, set up the project, make the change, test it, and prepare a pull request.
Beginners get to be a little lazier: no tab-hopping, no guessing, just the next small step. Slow and steady, one step at a time.
Demo
OpenPath runs locally with Ollama, so there is no hosted demo. [Add your video link or screenshots here.]
Code
aayushi2882
/
openpath
AI-powered onboarding for first-time open-source contributors
Stop searching. Start contributing.
OpenPath matches your skills with real GitHub issues, explains them in plain language with an AI that runs 100% on your own machine, and hands you a step-by-step roadmap to your first pull request.
✨ Features • 🧭 How it works • 🔒 Local AI • 🚀 Quick start • 🗺 Roadmap
😵 The problem
Almost every beginner wants to contribute to open source. Almost every beginner gets stuck on the same three questions:
🔍Thousands of issues, no idea which ones match your skills. |
🧩Maintainer jargon and missing context make issues unreadable. |
🚪Fork? Branch? Test? PR? Nobody tells you the order. |
OpenPath removes all three blockers — it takes you from finding an issue to understanding it to knowing exactly how to start.
✨ Features
🔗… |
How I Built It
The open-source AI at the core of OpenPath is Gemma (gemma4:e2b) running locally through Ollama.
- Frontend: React + Vite
- Backend: Node.js + Express
- Issue data: GitHub API
- AI: the backend talks to Ollama's local endpoint, where Gemma generates issue explanations and contribution roadmaps
- AI status check: the backend checks that Ollama is reachable and the model is installed, so setup problems show up before you try the AI features
Flow: repo URL + skills → Express fetches issues from GitHub → Gemma explains and plans locally → React shows recommendations, explanations, and roadmaps.
Why Does Open Innovation Matter?
OpenPath is built for students and beginners, and an open-weight model fits that audience:
- No API key, no bill. The local AI workflow needs no cloud AI key. Pull the model once and it's yours.
- Set it up once, use it forever. No usage limits or per-request cost while you're learning.
- Your questions stay on your machine. Beginners can ask "basic" questions about an issue without sending them to a hosted service.
- Small and capable enough. Gemma E2B is light enough to run on a regular machine and still explain issues clearly, which is all a beginner needs.
A closed API would have meant keys, quotas, and costs, exactly the friction OpenPath exists to remove.
The Lazy Payoff
Instead of opening a repo, reading issue after issue, researching what each one means, and planning the steps yourself, you paste one URL, enter your skills, and get a shortlist, a plain-language explanation, and a roadmap.
Next up (future scope): Playwright-based website inspection. Give OpenPath a website URL and it suggests UI/UX improvements, like cluttered cards or a missing footer, creating even more contribution ideas.
Prize Categories
Best Use of Gemma: OpenPath's whole AI workflow runs on Gemma (gemma4:e2b) locally through Ollama. Gemma generates the issue explanations and contribution roadmaps, and no hosted AI inference API is needed for that workflow.
Top comments (0)