Most running apps show you totals and leave you to interpret them. Run Log walks through one run and explains what changed, where, and what might account for it — with the confidence of every claim stated. Free, open source, and your file never leaves the browser.
No reviews yetBe the first to leave a review for Run Log
Maker
📌
Hey Product Hunt 👋
I kept finishing runs, opening the app, and looking at a slower third mile with no idea whether that was a hill, the heat, or me falling apart. The chart was right there. It just wouldn't say anything.
Run Log is my attempt at the missing half. You drop in a .fit or .gpx file from your watch and it walks through that one run in order, answering five questions:
What happened? — the totals, and what kind of run this was
Where did it happen? — a draggable timeline and a real map, synced to the same cursor
Which metrics changed together? — what moved at the same moment
What might explain it? — every explanation carrying its own confidence
What does that metric mean? — a short teaching note, specific to this run
The one rule I built it around: what your watch measured is kept visibly separate from what was inferred from it, and no inference can be written without stating how sure it is. That isn't a style guide — it's enforced in the type system. An explanation literally cannot compile without a confidence of high, medium or low, which the page renders as Likely explanation, Possible explanation, or Not enough data to be sure.
What it deliberately doesn't do: rank you, score you, or sort your splits fastest-first. A slow split is usually a hill, and a single number would hide exactly the context this thing exists to surface.
On privacy: your file is read in the browser and never uploaded. Exactly three things talk to a server, all listed in Settings, two of them off until you turn them on. The weather lookup rounds your coordinates to ~11km before they leave — and there's a unit test that exists solely to prove a precise position can't reach the URL, because that guarantee would otherwise fail silently.
It's free, MIT-licensed, has no account, and has no plans to become anything else. There's a demo run on the landing page if you want to look before you drop anything in.
I'd love to hear what you'd want explained that it doesn't explain yet. Happy to answer anything about the parsing, the widget architecture, or why I refused to build a leaderboard.