How AudioContext fingerprinting works
A page runs JavaScript that creates an AudioContext, often an OfflineAudioContext that doesn’t send sound to your speakers. The script then:
- Generates a known signal with an oscillator.
- Passes it through nodes such as a compressor or filter.
- Renders the signal and reads the resulting sample values.
- Converts selected values into a short hash.
Small differences can come from the browser’s audio engine, browser version, operating system, CPU architecture and floating-point calculations. Repeating the same test on the same software stack usually produces a similar result.
What the result means
The hash isn’t a device serial number and may be shared by many users. Its value comes from combining it with other browser fingerprinting signals such as screen size, installed fonts, time zone and graphics output.
Browser updates, operating system changes or privacy protections may alter the result. Some browsers limit the Web Audio API, standardize its output or add noise to reduce fingerprint stability.
AudioContext vs canvas fingerprinting
Both techniques generate content inside the browser and measure subtle rendering differences, but they test different subsystems.
| AudioContext fingerprinting | Canvas fingerprinting | |
|---|---|---|
| Subsystem | Audio processing | 2D text and graphics |
| Measured output | Audio sample values | Rendered pixel values |
| User permission | Usually not required | Usually not required |
Canvas fingerprinting is often paired with WebGL fingerprinting and audio measurements rather than used alone.
Why it matters with proxies
A proxy changes the IP address a website sees. It doesn’t change Web Audio output, so rotating IPs won’t rotate an AudioContext fingerprint. A stable fingerprint appearing from many IP addresses can become one input to bot detection, though responsible systems shouldn’t treat it as proof by itself.
For authorized automation and testing, use a consistent browser environment, keep request rates within the site’s limits and respect its terms and robots.txt. Site operators should disclose fingerprinting where required, collect only the signals they need and avoid making access decisions from one noisy measurement.
FAQ
Does AudioContext fingerprinting record my microphone?
No. The common technique generates and processes a synthetic signal inside the browser. Microphone recording uses a separate permission-controlled API.
Can a proxy prevent AudioContext fingerprinting?
No. A proxy changes network routing and the visible IP address, not the browser’s audio calculations. Browser privacy settings may restrict or standardize those calculations.
Is an AudioContext fingerprint unique?
Usually not by itself. Several devices can produce the same result, so websites often combine it with canvas, WebGL and other browser signals.