Launched this week

Melon Meter

Launched this week

Thump a watermelon. Your phone tells you if it's sweet.

7 followers

Thump a watermelon. Your phone tells you if it's sweet.

7 followers

Visit website

We all thump watermelons and nod like we heard something. It isn't folklore: ripe fruit is softer inside, so it rings lower — a gap of a few semitones, too small to hear. Thump 3 times, tap its rough size, get a 1-10. Optionally photograph the field spot and it reads that patch's exact colour and the rind's webbing. It measures ripeness, which correlates with sweetness — a phone, not a lab. Rate one after cutting and the scale re-fits to your melons. Free, no signup.

Free

Launch tags:Web App•Food & Drink•Physics

Launch Team / Built With

Framer AI AgentsDesign and publish professional sites with AI

Promoted

Maker

📌

Hey Product Hunt 👋 It's peak watermelon season, and most of us are still picking them the same way: thump it, nod like we heard something, buy it anyway. I wanted to know what we're actually supposed to be listening for. Turns out it's real physics. Ripe flesh is softer with bigger cavities inside, so the melon resonates lower. Unripe is dense and rings higher. The problem is the whole range is only a few semitones — too fine for our ears, easy for a microphone. So Melon Meter listens for you: Thump it three times. The mic catches each knock and picks out the melon's fundamental resonance. Tap its rough size. That mass-normalises the reading, so a giant melon isn't penalised for being big. Optionally, photograph the underside. Tap the pale field spot — the patch where it sat on the soil — and it reads that patch's exact colour, then scans the rest of the frame for the brown pollination webbing on the rind. Both nudge the score, and you see the arithmetic: thump, look, final. Two calls that made it work: Finding the right note. My first build used FFT peak-picking and kept locking onto the second harmonic, so every genuinely sweet melon scored a 1. Rebuilding around autocorrelation — which finds the fundamental rather than the loudest peak — was the difference between a novelty and something you'd trust. Cutting a feature. I'd originally asked people to report the rind's webbing themselves. Then I realised I couldn't reliably tell webbing from the melon's natural green stripes — and if I couldn't, nobody could. That question was feeding noise into the score, so I deleted it and taught the app to read it from the photo instead, where brown scarring and green stripes sit far apart in hue even though they fool the eye. Being straight about it: this measures ripeness, which correlates with sweetness — it's a phone, not a refractometer. It does get more personal over time though: rate four melons after you cut them and the scale re-fits to the fruit you actually buy. Free, no signup, nothing to install. It runs entirely in your browser — your audio and photo never leave your phone. Try it on your next melon and tell me how it did 🍉

Report

3d ago

love that it lets you rate after cutting so the scale actually learns your taste instead of pretending to know you from day one. the field spot + webbing combo is a clever touch too.

Report

2d ago

Maker

@safiyeacrt thank you! the rating loop was the part i cared about most, a score that doesn't learn is just a guess with a decimal. and yeah, field spot + webbing being two separate clues is exactly it.

Report

13h ago

finally an excuse to keep thumping melons at the store instead of pretending. tried it on a few in my backyard and the score lined up surprisingly well with the one i cut open.

Report

1d ago

Maker

@cafer_ak25358 this is the test i care about, thanks for cutting them open to check. "an excuse to keep thumping" is honestly the whole spirit of it.

Report

13h ago