PROOF CENTER

Proof organized around privacy, consistency, and protection

BotBrowser public validation already covers the model clearly: coverage, consistency standards, benchmark evidence, and rollout depth.

40+ validated scenarios4 supported platforms100+ isolated contextsOne browser model

Validation scope

Evidence is organized around signal dimensions, runtime paths, and public checks that show how the model stays aligned.

Consistency target

The same profile is expected to stay aligned across supported platforms and runtime modes.

Operating depth

The key parts of the browser model remain stable from first validation to scaled deployment.

How to read this evidence

Validation scope, benchmark coverage, and consistency standards define how the supported rollout path should be read.

VALIDATION METHODOLOGY

How the public validation model works

The public proof model is organized around scope, method, reproducible benchmarks, and alignment criteria.

01

What we validate

Signal groups, runtime paths, and platform combinations that represent the current public scope.

02

How we validate

Repeated public checks, internal replay targets, and cross-platform runs under the same browser model.

03

What counts as aligned

Stable profile output, no obvious runtime drift, and consistent behavior across the supported rollout path.

40+
Validated scenarios
Public checks plus internal replay targets
4
Supported platforms
Windows, macOS, Linux, Android
100+
Isolated contexts
Independent runtime model for scaled workloads
1
Browser model
The same core from first validation to enterprise rollout

PUBLIC BENCHMARKS

Public benchmark evidence under load and scale

Public benchmark results from the official repository show how the browser behaves under performance pressure and concurrent profile scale.

<1%
Speedometer overhead
Headed and headless runs stayed within run-to-run variance versus stock Chromium.
0ms
Added API latency
Canvas, WebGL, Navigator, Screen, and Font APIs showed no measurable added latency.
29%
Less memory at 50 profiles
Per-context architecture used less memory than launching 50 separate browser instances.
57%
Fewer processes at 50 profiles
Shared browser infrastructure cut process count at the same 50-profile scale.
01

Benchmarks were run locally across macOS, Linux, and Windows.

02

The official repository also reports roughly 2x faster profile creation at 50 concurrent profiles.

03

Public evidence here covers method, benchmark scope, and operating boundary.

VALIDATED COVERAGE

What is validated today

The signal dimensions and runtime paths below are the ones the product is built to keep aligned.

Signal collection coverage

Public proof includes the browser-facing signal groups most likely to reveal drift or leakage.

CanvasWebGLAudioUA-CH

Tracking correlation coverage

Validation also looks at where browser, network, and storage state start to correlate across runs.

Proxy pathSession continuityStorage stateIP linkage

Runtime parity coverage

The model is checked across runtime modes and host environments where output often starts to drift.

Headless parityGUI parityLinux runtimeAndroid targets

Identity control coverage

The proof surface also includes the layers that matter once the browser becomes part of a larger workflow.

FontsWebRTCGPU pathContext isolation

CONSISTENCY MODEL

Proof is about keeping the same model stable across checks and rollout stages

The strongest public proof point is that the same browser model stays coherent across platforms and rollout stages.

Same profile across platforms

The product is designed to keep one profile aligned across Windows, macOS, Linux, and Android targets.

Runtime discipline

Proxy paths, isolated contexts, and browser-level control stay part of the same deployment model.

Scale without model drift

The same proof model extends from early validation into long-running and enterprise-oriented deployment stages.

Map your target checks to the right validation path

Tell us which checks, platforms, and rollout stage matter to you, and we will walk through the model against your own targets.