Give every profile its own device-level characteristics. Two profiles open on the same Windows PC appear to fingerprinting services as two separate computers, so it is far harder to link an account in one profile to an account in another.
A website can identify your machine without cookies. A few lines of code reading your installed fonts, your graphics card, the way your computer renders text, the quirks of your audio hardware, and the precise timing of your clock are enough to build a stable identifier that follows you across sessions and accounts. That is fingerprinting, and it is how fraud-detection and bot services recognize a returning visitor even after cookies are cleared.
Incognito is not enough: it clears cookies but leaves the fingerprint identical. To keep account A from being linked to account B on the same hardware, each profile needs to look like a different device. That is exactly what this system does, and it works hand in hand with the profile system that keeps cookies, sessions, and history separate.
With protection on, MetaDock changes the answers your browser gives to dozens of identifying questions. The major categories:
| Signal | What MetaDock reports |
|---|---|
| Graphics card | A different GPU than you actually have, drawn from a library of 100+ real-world models matched to the operating system you are emulating |
| Screen | Resolution, color depth, pixel density, and taskbar size taken from real device profiles so the values stay internally consistent |
| CPU & memory | Core count and RAM amount from the same device profile (a budget laptop reports 4 cores / 8 GB; an enthusiast desktop reports 16 cores / 32 GB) |
| Canvas & text rendering | Each profile renders graphics with a small, consistent per-profile difference, so it looks like its own device |
| Audio | Each profile's audio fingerprint is shifted so it does not match your real hardware |
| Fonts | A platform-appropriate font set instead of your real installed-font list |
| User-agent & client hints | The browser/version/OS string and the modern Client Hints headers, set per profile |
| Network | Connection type and reported speed, controllable per profile |
| Secondary signals | Battery, plugins, permissions, and installed speech voices, normalized for cross-checking |
| Clock | High-resolution timing quantized so it does not leak machine-specific drift |
| IP & location | Your real IP and location, via the profile's proxy plus the geo system below |
You do not have to tune signals one by one. Pick a preset and MetaDock handles the rest, keeping the values coherent for you:
A profile that looks like a different machine on every visit is its own red flag. MetaDock seeds each profile so its spoofed values stay stable: the same profile produces the same fingerprint across restarts and weeks later. From a website's view, your profile is one consistent device.
If a profile's IP comes from London but it reports New York coordinates, a Berlin timezone, and Japanese as its language, fingerprinting services notice instantly. The geo system keeps these aligned:
These overrides apply before a page's scripts can read them, so a script asking for the timezone gets the spoofed answer. When geo spoofing is on, MetaDock also auto-approves the location-permission prompt so sites receive the spoofed coordinates without you clicking through a dialog. See the proxies guide for how to point a profile at a specific region.
The single biggest red flag is not any one value being unusual. It is values that contradict each other: a platform that says Windows paired with an Apple GPU, a user-agent claiming one OS while the platform setting says another, or a locale and timezone that point at different regions.
The Balanced and Maximum presets keep the whole bundle consistent automatically, which is why they are the safe default. If you switch to Custom and hand-edit individual signals, it is on you to keep the combination plausible: match the OS to the GPU, the user-agent to the platform, and the locale to the timezone.
It works for what it is designed for. MetaDock is built to make each profile look like its own ordinary device to the mainstream commercial fingerprinting techniques in wide use today. Profiles run independently without being linked to each other or to your real machine.