One of the most frustrating parts of software engineering isn't writing code—it's inheriting someone else's.
You clone a repository with thousands of files, dozens of folders, and years of history. Before you can fix a bug or add a feature, you have to answer questions like:
- Which files actually matter?
- Where is the real entry point?
- Which modules are safe to ignore?
- Where does the technical debt live?
- How is everything connected?
I've wasted enough time doing that manually that I decided to build a tool for it.
Introducing Project MRI
Project MRI analyzes a project ZIP and generates an onboarding guide for engineers. Instead of dumping you into a giant file tree, it tries to answer the questions you'd normally spend hours figuring out yourself.
It currently highlights:
- 📍 The first files you should read
- 🗺️ A high-level architecture overview
- ⚠️ Technical debt and risk hotspots
- 📊 A repository health score
- ⏱️ An estimate of how quickly someone can become productive
The goal isn't to replace your IDE or AI coding assistant. It's to solve the "where do I start?" problem.
Why I built it
Every time I joined a new project or came back to an old one, I found myself clicking through directories hoping to discover the important pieces.
I wanted something that could point me in the right direction immediately instead of making me reverse-engineer the repository from scratch.
Looking for honest feedback
I'm still actively improving it and would love input from engineers who work on large or legacy codebases.
Some questions I'm especially interested in:
- Would something like this actually save you time?
- What's the first insight you'd want after uploading a repository?
- What would make this genuinely useful in your day-to-day workflow?
If you'd like to try it, it's available here:
I'm not looking for compliments—I want to know what's confusing, missing, or not valuable enough to use.
Top comments (1)
Thanks for checking this out! I'm actively looking for honest feedback from engineers who work with large or legacy repositories. If something is confusing, inaccurate, or missing, I'd genuinely like to hear it.