← Back to Blog

How to Bypass Device Memory Fingerprinting: Protect Multi-Accounts with Proxies and Anti-Detect Tools

Device memory fingerprinting is a method of tracking based on the amount of RAM in a device. Proxies do not protect against this fingerprint, but there is a solution.

šŸ“…January 11, 2026
```html

You set up a proxy, use an anti-detect browser, but your accounts still get chain-banned? The problem may lie in Device memory fingerprinting — a tracking method that identifies your device by the amount of RAM. This parameter is transmitted to websites via the JavaScript API and remains the same for all your accounts, even if you change your IP address.

In this article, we will explore how device memory fingerprinting works, why regular proxies do not protect against it, and what tools can help arbitrageurs and SMM specialists avoid account linking.

What is Device Memory Fingerprinting

Device memory fingerprinting is a digital fingerprinting method that identifies a device based on the amount of installed RAM. Technically, this is implemented through the JavaScript API navigator.deviceMemory, which returns an approximate number of gigabytes of RAM in the system.

For example, if your computer has 16 GB of RAM, the API will return a value of 8 (rounded value). This value remains unchanged regardless of which browser you use, what IP address is set, or which cookies are cleared.

Why This is Dangerous for Multi-Accounting:

If you run 10 Facebook Ads accounts from different proxies but from one computer — they will all have the same deviceMemory value. Anti-fraud systems detect this match and link the accounts together, leading to chain bans.

Unlike an IP address, which can be easily changed via a proxy, the deviceMemory parameter is a characteristic of the physical device. Regularly changing proxies does not affect this value, so additional protection methods are required.

How Platforms Collect Memory Fingerprint

When you open a website or an ad account, JavaScript code runs in the background that collects dozens of parameters from your device. Device memory is just one element of a comprehensive digital fingerprint, but it is very stable and reliable for identification.

Here is what a simple script for collecting this parameter looks like:

// Example code executed by anti-fraud systems
const deviceMemory = navigator.deviceMemory || 'undefined';
const fingerprint = {
  memory: deviceMemory,
  cores: navigator.hardwareConcurrency,
  screen: `${screen.width}x${screen.height}`,
  timezone: Intl.DateTimeFormat().resolvedOptions().timeZone,
  // ... 50+ more parameters
};

// Sending data to the server for analysis
fetch('/track', { 
  method: 'POST', 
  body: JSON.stringify(fingerprint) 
});

Platforms collect this data each time you log into your account and compare it with previous sessions. If several different accounts show the same combination of deviceMemory + hardwareConcurrency (number of CPU cores) + screen resolution — the system marks them as linked.

Possible Values of navigator.deviceMemory

The API returns rounded values to protect user privacy, but this is not enough for anonymity in multi-accounting:

Actual RAM Size API Value Prevalence
≤ 0.25 GB 0.25 Very Rare (Old Devices)
0.5 - 1 GB 0.5 - 1 Budget Smartphones
2 - 4 GB 2 - 4 Mid-range Laptops, Smartphones
8 GB 8 Most Common Value
16 GB 8 Workstations of Arbitrageurs
32 GB and Above 8 Powerful Computers (Rare)

Note: if you have 16 GB or 32 GB of memory, the API will still return 8. This makes the fingerprint less unique but does not solve the account linking problem — because in combination with other parameters (number of cores, GPU, screen resolution), the device is still identifiable.

Why Proxies Are Not Enough to Protect Against This Method

Many novice arbitrageurs think: "I bought residential proxies, each account has its own IP — so I am protected." This is a dangerous misconception that leads to the loss of accounts and advertising budgets.

A proxy server operates at the network level — it replaces your IP address and routes traffic through another server. But JavaScript code runs locally in your browser and has access to the characteristics of your physical device. Proxies do not affect the values returned by navigator.deviceMemory.

What Proxies Protect vs What Remains Vulnerable

Parameter Do Proxies Protect? Comment
IP Address āœ“ Yes Main function of proxies
Geolocation (by IP) āœ“ Yes Shows the location of the proxy server
WebRTC Leaks ~ Partially Requires additional browser configuration
Device Memory āœ— No Read locally via JS
Hardware Concurrency (CPU Cores) āœ— No Local device parameter
Canvas Fingerprint āœ— No Depends on GPU and drivers
User-Agent āœ— No Sent by the browser
Screen Resolution āœ— No Physical monitor parameter

As you can see, proxies only solve one task — IP address replacement. All other parameters of the digital fingerprint remain unchanged and reveal your real device.

Real Case:

An arbitrageur launched 15 Facebook Ads accounts through quality residential proxies. Each account had a unique IP from different cities in the USA. After 3 days, all accounts were banned simultaneously. The reason: identical deviceMemory values (8 GB) + hardwareConcurrency (12 cores) + screen resolution 1920x1080. The system determined that all accounts were running from one physical computer.

Which Platforms Use Device Memory Fingerprinting

Device memory fingerprinting is a standard part of modern anti-fraud systems. Almost all major platforms that combat multi-accounting and bots collect this parameter. Here is a list of platforms where this method is definitely used:

Advertising Platforms (Critically Important for Arbitrageurs)

  • Facebook Ads / Meta Business Suite — one of the most aggressive fingerprinting systems. Collects over 100 device parameters, including deviceMemory, and actively uses them to link accounts.
  • Google Ads — uses comprehensive fingerprinting through Google Analytics and its own scripts. Device memory is part of the parameters for detecting suspicious activity.
  • TikTok Ads — actively develops anti-fraud systems and uses fingerprinting to combat account farming. Particularly sensitive to matching hardware parameters.
  • Yandex.Direct — the Russian platform has also implemented fingerprinting to protect against click fraud and multi-accounting.

Social Networks (Important for SMM Specialists)

  • Instagram — as part of the Meta ecosystem, uses the same fingerprinting technologies as Facebook. Collects deviceMemory during registration and each login.
  • TikTok — besides the ad cabinet, the social network itself actively combats inflation and bots through fingerprinting.
  • LinkedIn — used to protect against spam accounts and automated data scraping.
  • Twitter/X — collects hardware parameters to combat bots and coordinated attacks.

E-commerce and Marketplaces

  • Amazon — uses fingerprinting to protect against fraud and multi-accounting of sellers. Linking accounts can lead to the blocking of all stores.
  • eBay — collects device fingerprint during registration and product listings.
  • Wildberries, Ozon — Russian marketplaces are also implementing fingerprinting technologies to combat review inflation and multi-accounting.

Financial Services and Payment Systems

  • PayPal — one of the pioneers in using fingerprinting to protect against fraud.
  • Stripe — collects an extended set of device parameters to assess transaction risks.
  • Banking Apps — almost all use fingerprinting as an additional authentication factor.

It is important to understand that the deviceMemory parameter alone does not lead to bans. Platforms analyze the combination of dozens of parameters and look for suspicious patterns. For example, if 10 different accounts have the same deviceMemory + hardwareConcurrency + canvas fingerprint, but different IP addresses — this is a clear sign of multi-accounting.

How Anti-Detect Browsers Spoof Memory Values

Anti-detect browsers are specialized tools that allow you to create isolated profiles with unique digital fingerprints. Unlike regular browsers, they can spoof JavaScript API values, including navigator.deviceMemory.

Popular anti-detect browsers for multi-accounting:

  • Dolphin Anty — one of the most popular among Russian-speaking arbitrageurs. Free plan for 10 profiles, user-friendly interface, quality fingerprint spoofing.
  • AdsPower — a powerful tool with automation and API. Suitable for large operations with hundreds of accounts.
  • Multilogin — premium solution with its own browser engines Mimic (Chromium) and Stealthfox (Firefox). The highest quality fingerprint spoofing, but expensive.
  • GoLogin — an affordable alternative with good price/quality ratio.
  • Octo Browser — a new player in the market, actively developing, focusing on teamwork.

DeviceMemory Spoofing Technology

Anti-detect browsers work at the JavaScript engine level and intercept calls to the browser API. When a website tries to read navigator.deviceMemory, the browser returns not the actual value of your device, but a spoofed one — the one you specified in the profile settings.

Simplified, it works like this:

// Actual value of your device
console.log(navigator.deviceMemory); // 8 (16 GB RAM)

// After spoofing by the anti-detect browser
Object.defineProperty(navigator, 'deviceMemory', {
  get: function() { return 4; } // Spoofed value
});

console.log(navigator.deviceMemory); // 4 (spoofing active)

It is important that anti-detect browsers spoof not only deviceMemory but also related parameters:

  • hardwareConcurrency — number of logical CPU cores. If you set deviceMemory = 4, the browser will automatically select a realistic core value (for example, 4 or 8).
  • Canvas Fingerprint — unique GPU fingerprint. Generated randomly or taken from a database of real devices.
  • WebGL Parameters — information about the graphics card and drivers.
  • User-Agent — browser identification string, must match the spoofed characteristics.

Important — Consistency of Fingerprint:

You cannot just specify a random deviceMemory value. All parameters must match each other. For example, if you set deviceMemory = 2 (weak device), but hardwareConcurrency = 16 (powerful processor) — this looks suspicious. Good anti-detect browsers automatically monitor the consistency of parameters.

Step-by-Step Setup of Protection in Dolphin Anty and AdsPower

Now let's move on to practice. We will discuss how to properly set up deviceMemory spoofing in the two most popular anti-detect browsers among arbitrageurs and SMM specialists.

Setup in Dolphin Anty

Dolphin Anty is a free anti-detect browser with a limit of 10 profiles. This is sufficient for most tasks. Here is a step-by-step guide:

Step 1. Create a New Profile

  1. Open Dolphin Anty and click the "Create Profile" button
  2. Enter the profile name (for example, "FB Ads - Account 1")
  3. Select the platform — specify for which social network or advertising platform the profile is created (Facebook, TikTok, Google, etc.)

Step 2. Configure Fingerprint

  1. Go to the "Fingerprint" tab
  2. Generation mode: select "Real" — the browser will take parameters from a database of real devices
  3. Operating system: select Windows 10 or macOS (depending on your strategy)
  4. Find the "Device memory" parameter — set a value different from your actual one. Recommendations:
    • For mobile accounts: 2-4 GB
    • For desktop accounts: 4 or 8 GB
    • Avoid too rare values (0.25, 0.5)
  5. Hardware concurrency (cores): will be set automatically according to deviceMemory. If needed, you can change it manually (4-8 cores — realistic)

Step 3. Configure Proxy

  1. Go to the "Proxy" tab
  2. Select proxy type: for Facebook Ads and Instagram, mobile proxies are recommended, for Google Ads and TikTok, residential proxies will suffice
  3. Enter proxy details in the format: IP:Port:Username:Password
  4. Click "Check Proxy" — make sure the connection works and geolocation is determined correctly
  5. Important: the geolocation of the proxy must match the timezone in the fingerprint settings

Step 4. Additional Settings

  1. WebRTC: set to "Altered" — this will prevent your real IP from leaking through WebRTC
  2. Canvas: "Noise" mode — adds random distortions to the canvas fingerprint
  3. WebGL: "Noise" or "Spoof" mode — protection against identifying the real graphics card
  4. Geolocation: specify the coordinates of the city that corresponds to the proxy's geolocation

Step 5. Save and Launch

  1. Click "Create Profile"
  2. The profile will appear in the list — click "Launch"
  3. A browser with a spoofed fingerprint will open
  4. Check the settings on the site browserleaks.com or pixelscan.net

Setup in AdsPower

AdsPower is a more advanced tool with automation and API. The setup process is similar to Dolphin Anty, but with some differences:

Step 1. Create Profile

  1. Click "New Profile" in the main window
  2. Select profile type: "Quick Setup" or "Custom"
  3. Custom is recommended for full control over the fingerprint

Step 2. Fingerprint Settings

  1. Operating System: select Windows or macOS
  2. User-Agent: select "Automatic" — it will be generated automatically based on the OS
  3. Hardware Settings:
    • Device Memory: set 2, 4, or 8 GB (avoid matching your real value)
    • CPU Cores: AdsPower will automatically select a realistic value, but you can change it manually
  4. Screen Resolution: choose a resolution corresponding to the selected device type (for example, 1920x1080 for desktop, 1280x720 for tablet)

Step 3. Proxy Configuration

  1. Go to the "Proxy Settings" tab
  2. Proxy Type: select HTTP, HTTPS, or SOCKS5 (depending on your proxy provider)
  3. Enter details: Host, Port, Username, Password
  4. Click "Check Proxy" — AdsPower will show IP, country, and connection speed
  5. If you are using IP rotation, set the rotation interval (for residential proxies, this is usually not required)

Step 4. Advanced Settings

  1. WebRTC: set to "Alter" — replaces local IP with proxy IP
  2. Canvas: "Noise" mode — adds noise to the canvas fingerprint
  3. WebGL: "Noise" — protection against GPU identification
  4. Fonts: AdsPower automatically selects a font set according to the OS
  5. Timezone: must match the geolocation of the proxy (AdsPower can set it automatically)

Step 5. Save and Test

  1. Click "Save" to save the profile
  2. Launch the profile by clicking "Open"
  3. Check the fingerprint on verification services (described in the next section)

Comprehensive Protection: Proxy + Fingerprint Spoofing

Spoofing deviceMemory is just one element of protection. For reliable multi-accounting, a comprehensive approach is needed, which includes the right choice of proxies, configuration of the anti-detect browser, and adherence to operational security.

Rule 1: One Profile = One Proxy = Unique Fingerprint

Never use the same proxy for different accounts that should not be linked. Each profile in the anti-detect browser must have:

  • A unique IP address (via proxy)
  • A unique set of fingerprint parameters (deviceMemory, hardwareConcurrency, canvas, etc.)
  • Separate cookies and localStorage
  • Different User-Agent strings

Rule 2: Choose Proxy Type Depending on the Platform

Not all proxies are equally suitable for different tasks. Here are recommendations for selection:

Platform Recommended Proxy Type Why
Facebook Ads, Instagram Mobile Proxies Maximum trust, IPs from mobile operators rarely get blacklisted
TikTok Ads, TikTok Mobile or Residential The platform is aimed at mobile users
Google Ads Residential Proxies Good balance of price and quality, stable IPs
LinkedIn, Twitter/X Residential Proxies Average level of protection, residential proxies are sufficient
Amazon, eBay Residential (Static) Stability of IP is needed, frequent changes raise suspicions
Parsing Marketplaces Datacenter Proxies High speed, low price, suitable for parsing

Rule 3: Data Consistency

All profile parameters must be logically related:

  • Proxy Geolocation = Timezone = Language. If the proxy is from Los Angeles, the timezone should be America/Los_Angeles, and the browser language — en-US.
  • DeviceMemory corresponds to the device's power. If you specified 2 GB of memory, do not set 16 CPU cores — this is unrealistic.
  • User-Agent corresponds to the OS and resolution. If you chose macOS, the User-Agent should contain "Macintosh", not "Windows NT".
  • Canvas and WebGL fingerprints correspond to the specified graphics card. Good anti-detect browsers do this automatically.

Rule 4: Account Warming

Even with a perfect fingerprint and quality proxies, a new account can get banned if aggressive actions are started immediately. Recommendations for warming up:

  1. Day 1-3: Registration, profile filling, adding photos. No advertising actions.
  2. Day 4-7: Likes, comments, content viewing. Imitation of a regular user.
  3. Day 8-14: First posts, adding friends/followers.
  4. Day 15+: You can start advertising activities, but gradually increasing the volumes.

Important: always log into the account through the same proxy. Changing IP for an already registered account is a red flag for anti-fraud systems.

Rule 5: Regular Fingerprint Rotation (for New Accounts)

If you are constantly registering new accounts, do not use the same combinations of deviceMemory + hardwareConcurrency + canvas. Create diversity:

  • Alternate deviceMemory values: 2, 4, 8 GB
  • Change the number of cores: 4, 6, 8, 12
  • Use different screen resolutions: 1920x1080, 1366x768, 1440x900
  • Alternate OS: Windows 10, Windows 11, macOS

Anti-detect browsers often have a "Random fingerprint" or "Generate from real devices" feature — use it for automatic creation of realistic combinations.

How to Check Your Memory Fingerprint

Before launching advertising campaigns or registering accounts, be sure to check the quality of fingerprint spoofing. There are several specialized services for this.

Services for Checking Fingerprint

1. BrowserLeaks.com

The most popular and detailed service. Provides detailed information about your digital fingerprint.

What to check:

  • Open browserleaks.com/javascript
  • Find the "navigator.deviceMemory" parameter — check that the value matches the settings in the anti-detect browser
  • Check "navigator.hardwareConcurrency" — the number of cores should be realistic for the specified memory
  • Open the "WebRTC" section — make sure your real IP is not leaking
  • Check Canvas fingerprint — it should differ for each profile

2. PixelScan.net

Another service for checking your fingerprint. Provides similar information and is user-friendly.

Follow the same steps as with BrowserLeaks.com to ensure your fingerprint is properly spoofed.

Conclusion

Device memory fingerprinting is a significant challenge for multi-accounting and arbitrage. Regular proxies are not enough to protect against this method. Using anti-detect browsers and following best practices can help mitigate the risks and keep your accounts safe.

```