Every developer knows the feeling. New job, new repo, 100k+ lines of code, and the person who wrote it left months ago. The README is three commands, all outdated. You spend your first two weeks just trying to figure out where things live.
I got tired of that, so I built a tool to fix it.
The problem with reading code top-to-bottom
When you open an unfamiliar codebase, the instinct is to start reading files. But you don't know which files matter. You waste hours in utility folders and config that teach you nothing about how the system actually works.
The fastest engineers I know do the opposite: they find the 5 files that carry the most meaning and read those first, in order. Entry point, then core domain logic, then the riskiest / most-changed files.
Project MRI does this automatically
Point it at a repo and it gives you:
- The 5 files to read first — in the right order
- An architecture diagram — how the pieces connect
- A risk radar — the most fragile, most-changed files that break things
- A health score — how much technical debt you're inheriting
No signup for the demo. Try it on any public repo: https://projectmri.com
Try it on a repo you already know
The best test: run it on a codebase you know cold. Does the "read these first" order match how you'd actually onboard a new teammate? That's the bar.
If you inherit codebases for a living, the full version is $99: https://projectmri.com
Would genuinely love feedback on where the file-ordering gets it wrong — that's what I'm tuning right now.
Top comments (0)