Quiet Meter

Overview

Quiet Meter is an iPhone sound level meter. It shows a live reading in numerals you can read across a room, the minimum, average and maximum for the session, a rolling 45-second chart coloured by level, and a reference table that says what 40, 60, 80 and 110 dB actually sound like.

It exists because of what the category’s one-star reviews say. Every popular decibel app on the store is free to download and then asks for $40 or $50 a year for a tool that reads a microphone. Quiet Meter is free with an occasional full-screen ad, and one $4.99 payment removes the ads and opens the full session history. There is no subscription in it and there never will be.

Being Honest About What It Measures

An iPhone microphone is designed for speech. It applies automatic gain, compresses very loud sounds and rolls off deep bass. Whatever the app puts on screen is an estimate, not a measurement.

That constraint shaped the product rather than being disclosed at the bottom of a settings page. Every surface says it, the caveat is embedded in every exported session, and there is no OSHA reference, no NIOSH limit, no exposure timer and no hearing-damage claim anywhere in the app or the listing. Dropping those was a deliberate trade: they are the most searched-for features in the category, and they are also what pulls an app into medical-adjacent territory it cannot honestly occupy.

The reference table carries no exposure limits for the same reason. It says a library is about 40 and a concert is about 110, which is the useful part.

Calibration That Survives the Session

The most common complaint about decibel apps is that they read 28 dB in a silent room and there is nothing you can do about it. Quiet Meter lets you set an offset from −20 to +20 dB in half-decibel steps against a meter you trust, and it stays set.

The interesting part is what that does to a measurement already in progress. Session statistics accumulate in raw dBFS rather than in displayed decibels, so the offset is applied at the point of display. Recalibrate halfway through and the minimum, average and maximum all shift together — you get one corrected session rather than two spliced scales.

The chart is drawn as plain React Native views, downsampled to a fixed bar count. No chart library, no SVG dependency, nothing to keep up to date.

Pricing as a Screenshot

The meter, the statistics, the chart and the whole calibration screen are free and never gated. The free tier shows the newest three saved sessions; older ones are hidden rather than deleted, and buying Pro brings them back exactly as they were saved. Advertising appears only after a measurement is saved or shared — never while the meter is running, never while the microphone is on, and not at all in the first sessions.

Privacy by Construction

The app has no backend, no account, no analytics and no network code of its own. Sessions live in AsyncStorage on the device.

The microphone is a separate story and it is the one worth getting right. iOS has no way to read a level without a recorder, so a recorder runs — writing to the app’s own cache container, never read, never copied, never uploaded, torn down the moment the meter stops or the app is backgrounded. The advertising module is deliberately built so it cannot learn anything from it: it exposes only zero-argument functions, and the placement decision takes a handful of integers — a saved-session count, a break count, two timestamps — and no reading, session, note or recording.

Ads are non-personalised, the app never presents an App Tracking Transparency prompt, and the privacy manifest declares no tracking. Those four facts are consistent with each other by design; changing one would require changing all of them.

Current State

Quiet Meter 1.0 is with App Review, submitted on 9 September 2026 with its listing, both screenshot sizes and the one-time Pro purchase attached.

Before it went, the interface was swept across six routes, two device sizes, light and dark, and fresh and seeded states — 48 combinations — and driven through nine flows headlessly. That found four real defects: a paywall whose five points would not fit a 6.7“ screen, a meter screen that pushed its accuracy note off the bottom, and two rows of tap targets under the 44-point minimum.

The one thing a simulator cannot answer is whether the reference constant that converts microphone level to estimated dB SPL is a sane default, because a simulator has no usable microphone input. That is the first thing to check against a real meter on real hardware.