Skip to content
Glossary

WebGL fingerprinting

A fingerprinting technique that reads a device's GPU vendor and renderer strings and hashes rendered 3D output to help identify the browser.

WebGL fingerprinting extracts identifying information from the browser's 3D graphics stack. It has two parts. First, the WebGL API exposes the unmasked vendor and renderer strings, which name the actual GPU, for example a specific Apple, NVIDIA, or Intel part. Second, scripts can render a 3D scene off-screen and hash the resulting pixels, which vary with the GPU, driver version, and operating system. Together these reveal hardware details that narrow a device down considerably.

The GPU strings are especially significant for anyone presenting a configured browser identity, because they must agree with everything else the browser claims. A profile that reports Windows in its user agent but exposes an Apple GPU renderer string is internally inconsistent, and consistency checks like this are exactly what fingerprinting scripts look for. The same applies in QA and ad verification work, where a test environment should faithfully resemble the configuration being tested.

Oculr controls the WebGL unmasked vendor and renderer strings at the engine level, exposed through native getters that return the same native code signatures as stock Chrome, and applies seeded, deterministic per-profile noise to WebGL image output. GPU strings derive from the same single profile definition as the user agent, platform, and client hints, so the identity holds together across surfaces.

Real engine
Fingerprinting compiled in
20+ kernels
Chrome 86 to current majors
40+
MCP agent tools

Give your agents a browser that sites trust

Spin up your first two profiles free. Point an agent at them, run a task, and watch Oculr turn it into a workflow you can scale.

No credit card required · Cancel anytime