How site separation protects your browser.

1. Threat Model: Silent Audio Fingerprinting

The Web Audio API can process a silent signal without playing sound. Small differences in hardware, operating systems, browsers, and audio processing may produce a repeatable result that becomes one part of a browser fingerprint.

A tracking script can measure that result and combine it with other signals to help recognize a browser after cookies are cleared or across unrelated websites.

SoundPrint Guard's objective is focused: prevent unrelated websites from receiving the same reusable Web Audio identity while keeping results consistent enough for normal site behavior.

2. Why Naive Random Noise Fails

Older extensions attempt to defend against audio fingerprinting by injecting random Math.random() noise on every API call. This approach introduces fatal security and reliability flaws:

• Instability can look suspicious: Anti-abuse systems may notice values that change unpredictably between consecutive reads.
• Cross-API Incoherence: If a script calls getChannelData() and copyFromChannel(), random noise causes the two methods to return conflicting sample values, easily detected by anti-tamper heuristics.
• Audio Distortion: Unbounded noise in live playback buffers causes audible clicks, pops, and breaks web synthesizers.

3. Site-Specific Identity Isolation

SoundPrint Guard creates a private key during installation and keeps it in extension-controlled browser storage. That local key is used to produce a separate identity for each website without sending the key, browsing history, or audio content to a server.

Core design properties:

• Same-site consistency: Reloads and parallel tabs for the same website receive a coherent identity.
• Cross-site separation: Unrelated websites receive different identities, preventing simple reuse of the same audio result.
• Related API consistency: Browser methods that describe the same audio data are kept aligned instead of returning contradictory random values.

4. Seven-Surface Protection Health

SoundPrint Guard checks the major Web Audio paths commonly used to build an audio fingerprint. The extension summarizes these checks as Protected, Limited, or Needs attention.

5. Local Activity Classification

A local classifier looks for combinations of behavior associated with audio fingerprint extraction while accounting for ordinary audio use in games, calls, media players, and browser-based studios.

• Context matters: One audio operation is not automatically classified as tracking.
• Combined behavior matters: Silent rendering followed by immediate, repeated analysis is more significant than normal playback.
• Classification is not accusation: A warning describes technical similarity to fingerprinting behavior, not the intent or legal status of a website.

6. Browser Surface Compatibility

The protection layer is designed to preserve the normal shape and behavior of browser methods so compatibility checks do not encounter obvious contradictions.

• Coherent results: Related audio methods remain consistent with one another.
• Expected method behavior: Protected methods retain the browser-facing characteristics required by normal page code.
• Compatibility controls: Balanced, Maximum, Compatibility, and per-site Disabled modes provide clear fallbacks.

7. Scope and Boundaries

SoundPrint Guard specializes strictly in Web Audio fingerprint isolation. A VPN hides or replaces the public IP a website sees, while private browsing limits local history and temporary cookies. Neither normally changes the audio characteristics exposed by the same browser and device.

These protections are complementary: use SoundPrint Guard for Web Audio fingerprint separation, a VPN or Tor for network anonymity, and private browsing when you want a temporary local session. SoundPrint Guard does not replace a VPN, ad blocker, or protection for other fingerprinting surfaces.