Launched this week
PulseRepo
Turn any repo's git history into questions to ask
9 followers
Turn any repo's git history into questions to ask
9 followers
PulseRepo reads the process behind a codebase, not the code. It turns git history into plain-language questions - who owns what, what's stuck, whether tests keep up - for the people accountable for the code, not the ones writing it. Never scores people. Free public tool, no signup.
Love the angle of focusing on ownership and flow rather than lines of code. One thing that would make it stickier for me: letting me save a watchlist of repos so I can get a weekly digest of who's stuck and what's drifting. Right now each visit feels like a one off check.
@sudenurouou Thank you - that ownership-and-flow lens is exactly what I hoped would land, so this genuinely means a lot.
You have put your finger on the real gap. The free /scan is deliberately a one-off, anonymous check (no signup, nothing stored), so by design it can't remember anything between visits. But that "track it over time" need is the whole point of the signed-in side: it already keeps history and trends across multiple repos, so you can watch drift build up instead of re-checking cold each time.
A recurring "here is who got more stuck and what drifted since last week" digest is the natural next layer - and honestly not shipped yet, it's on the roadmap, not live. Your framing of it as a watchlist is cleaner than how I had it in my head, so I'm stealing it.
If you did get that weekly, what would you want at the top: the new risks that just appeared, or the deltas on the ones you already flagged?
the choice to frame questions around ownership and stuck work instead of individual output feels really considered. no signup for a no-login tool is a small thing but it makes me trust the product before i even read the results.
@sonerazv7 Thank you - you named the two decisions I went back and forth on the most, so it means a lot that they read as trust rather than as gaps.
Both were deliberate. Never grading individual output isn't something I'll "fix" later - it's a hard line: the moment a tool like this scores a person, it becomes something that gets weaponized in a review, and the signal is worthless anyway. And the no-login was the same instinct you're describing: if I'm asking you to trust a read on your repo, making you hand over an account first is exactly the wrong opening move.
The real test now is whether the results live up to that first impression. If you paste a repo, I'd love to know whether any single question feels off - especially one that reads like it's judging a person instead of the work. That's the feedback I'm hunting for today.
Finally tried this on our repo and the ownership breakdown actually matched what I knew was happening, which was not what I expected from a free tool. The no-scoring thing feels like the right call too.
@harundoantszjt Thank you - that's the bar I care about most: the ownership read has to match what an insider already knows, or nothing built on top of it is trustworthy. Really glad it held up on a real repo.
The "free tool" surprise is the deterministic part doing its job - ownership, coupling, solo-owned code are plain git math over your history, not a model guessing, so there is nothing to hallucinate. That is also what let me make it free and no-login.
Since the breakdown matched what you already knew, I would love your read on the harder half: did any question surface something you did NOT already know, or was it mostly confirming what was already in your head? The confirmation is table stakes - the moment it tells an accountable person something they would have missed is the whole point, and that is the part I am least sure lands yet.