Saltar al contenido

Esta página aún no está en su idioma. Está leyendo el texto en inglés.

Free tools

Audio fingerprint

A site runs a tone through a compressor with no speaker output and sums the samples. The sum is the audio fingerprint, and here yours is shown together with whether it repeats.

Your values right now

The tone is computed on this page without a single sound from the speakers and goes nowhere.

Audio number
Midiendo…
Two runs in a row matched
Midiendo…

Todo en esta página se mide en tu navegador y nunca sale de él. Sin peticiones, sin registros, sin cuenta.

How the number is computed

The page creates an audio context with no output, passes a tone through it and reads the resulting samples. Different audio paths give slightly different samples, and the sum comes out particular to the machine.

Why silence does not count

Reading an empty analyser gives the same silence on any browser, with or without substitution. Sites read audio through a compressor, and this check does the same.

What ARMANOS does

The audio shift is derived from the profile number and does not change between launches. Two profiles of one build already differ in the seventh decimal place, and that is enough for a site.

The Fingerprint self-test inside the program (profile menu > Test fingerprint) reads audio the way a site does: through a compressor, not through an empty analyser.

How to compare two profiles

Open the page in two profiles. The audio numbers have to differ, and the «two runs in a row matched» row has to answer yes in both.

What this check does not show

  • The number on other sites is different: they have their own tone and their own run length. What matters is that two profiles differ and one profile repeats.
  • Browsers that keep audio closed until a click. Their first row answers empty, and that is a property of the browser, not a fault.
  • Microphones and speakers. Their count is a separate reading, and it is not read here.

Where this is decided

The readings are measured in your browser, and the promises below are backed by files you can open.

Readings on the page come from an allowed list
apps/web/components/site/LiveReadout.tsx
The Fingerprint self-test reads audio through a compressor, like a site
apps/desktop/test/audit-on-builtin.js
A profile's audio is stable between launches and differs between two profiles
apps/desktop/test/fingerprint-consistency.js

Questions

Why do I hear nothing?
The run goes without speaker output. Sites compute audio the same way and need no speakers.
What does «no» in the second row mean?
Two runs of one tone gave different sums. That does not happen on a live machine.
Is anything sent out?
No. The tone is computed in your browser.

Two profiles give two different answers

In ARMANOS these values are decided inside the browser engine before the page gets anything. Open two profiles, run the check in both and compare.