Audio Fingerprint
Your audio fingerprint uniqueness
Estimating…
Computed locally in your browser. Uniqueness is a heuristic estimate — not from a global database.
How to use
- Open the page — audio probe runs automatically (may take a second).
- Read the sample sum and fingerprint ID in the headline.
- Refresh after browser updates or privacy mode changes.
FAQ
What is audio fingerprinting?
Scripts render audio offline through OscillatorNode and DynamicsCompressor, then read sample output. Floating-point and DSP differences across OS and hardware produce distinct sums/hashes.
Why unavailable or blocked?
Some browsers disable OfflineAudioContext in private mode or return normalized output anti-fingerprinting. Safari and hardened profiles may differ from Chrome.
Do I need to allow microphone?
No. This uses OfflineAudioContext — no mic permission. It is not recording your room.
Will my ID match other audio fingerprint sites?
Probe parameters (frequency, compressor settings, sample range) affect the sum. This tool uses a fixed educational recipe; commercial trackers may differ.
Same technique as trackers?
Same family — oscillator + compressor + sample read — described in academic and industry fingerprint literature.
Is audio played aloud?
No. Rendering is offline into a buffer; you should not hear anything.
Introduction
Audio Fingerprint implements a classic OfflineAudioContext probe: triangle oscillator at 10 kHz through a dynamics compressor, render offline, sum absolute sample values in a fixed index range, then hash the result. Trackers use similar tricks because audio DSP paths vary subtly by OS, hardware, and browser build.
When browsers mitigate audio fingerprinting, you may see blocked or a fixed sum — both outcomes are informative for privacy reviews.
Probe recipe
| Step | Setting |
|---|---|
| Context | 1 channel, 44100 samples, 44100 Hz |
| Oscillator | Triangle wave, 10000 Hz |
| Compressor | threshold −50, knee 40, ratio 12 |
| Readout | Sum of abs(samples[4500..4999]) |
The headline ID hashes the sum string with SHA-256 (first 16 hex chars shown).
Common use cases
- Privacy extension audit — see if audio output is randomized.
- Cross-browser lab — compare Chrome vs Firefox vs Safari sums on one machine.
- Complete fingerprint stack — add canvas and WebGL layers.
- Research notes — document audio layer separately from UA hash (browser fingerprint).
Best practices
- Run with sound muted or unmuted — offline render should not depend on speakers, but browser bugs vary.
- Retry once if the first load timed out on a slow device.
- Do not confuse with mic test — microphone tools use live input, not offline synthesis.
- Full combined ID: what is my fingerprint.