Home
Short updates about CrowdMe.One.
Starting a CrowdMe.One session is now faster.
After a normal login, your computer now opens automatically if it is ready. No need to click Open Computer first.
For an even faster start, you can bookmark or add a link to your desktop: https://crowdme.one/login?token=XXXX
Replace XXXX with your token and the link will take you directly into a new CrowdMe session.
Note: Only save this link on a private device you trust. The link contains your CrowdMe token.
Tor Browser is now available as an option in CrowdMe.One.
In our fingerprint tests, all measured groups remain identical between CrowdMe sessions except Canvas. Tor Browser deliberately randomises the Canvas result, and in our tests it changes even after refreshing the same site.
This makes the underlying CrowdMe environment stable, while Canvas remains intentionally different on every test.
To use Tor Browser, unmaximize Firefox and start Tor Browser from the left-side toolbar.
We've been testing what browser data can be kept identical between CrowdMe sessions.
Our test at crowdme.one/testfp.html currently checks around 150 browser-visible parameters. In our CrowdMe environment, all tested values can be kept stable between sessions — with one interesting exception: Firefox's Canvas protection.
Firefox deliberately randomises the Canvas result. In our testing, that random value is stable for a site during a Firefox session, but changes between sites and between Firefox sessions.
This means the randomised value can itself act as a temporary identifier. A site can potentially recognise repeated visits during the same Firefox session, even without cookies. It does not, however, provide the same identifier across different sites or new Firefox sessions.
Is that better than returning the same Canvas result as every other CrowdMe user? In our particular environment, perhaps not. But for an ordinary browser, the trade-off is different.
With CrowdMe, you can always request a new computer. This removes not only this temporary Canvas identity, but cookies, browser state and other session-specific identifiers — including ones we may not know about.
The useful finding isn't that randomisation is good or bad. It's understanding exactly what identity it creates, how long that identity lasts, and where it can be used.
Sound can now be enabled from the session menu. Playback may have a few seconds of delay depending on browser buffering.
CrowdMe normally gives everyone the same fingerprint. RandomizeFP lets you experiment by randomising hardware, language and time zone, combined with different screen resolutions and orientations.
After months of development and testing, CrowdMe.One is now live and available for anyone to use.
Recent additions include more than 40 proxy exit countries alongside Tor, improved copy and paste, and clearer documentation of how CrowdMe works, its threat model, and the browser fingerprint problem it is designed to address.
We've also published two new articles: The Fingerprint Dilemma and Beyond Cookies.
And since we still like as much testing and feedback as possible, these tokens each contain 100 free sessions. First come, first served:
Q238-R893-E362-N853
F874-T725-G453-Z675
S626-M559-C767-R783
S747-J368-X953-E544
P846-X839-C762-U278
A774-C758-T848-H384
C467-H754-V872-Z499
E345-D649-N933-X724
D498-K229-P373-X766
G384-G298-E242-V897
Try it at CrowdMe.One.
How should browser fingerprint protection actually be evaluated?
Our latest research note explores two fundamentally different approaches to browser fingerprint protection and asks a simple question: how do we know that it really works?
Cookies are no longer the only—or even the biggest—privacy challenge.
We've published a short article explaining current techniques, including practical examples and links to browser fingerprinting tools that you can try yourself.
Copying and pasting between your local computer and the remote browser is now much easier. New Copy and Paste options are available directly from the session menu.
The new functions eliminate the need to use the noVNC clipboard panel for everyday clipboard operations, making file and text transfers faster and more convenient.
CrowdMe.One now includes a Firefox extension that lets you quickly choose the Tor exit country used by your browser session.
Click the CrowdMe.One icon in the Firefox toolbar and select an exit country from the list. New connections will use the selected country immediately. Existing connections may continue to use the previous exit until they are reloaded.
Available options currently include Random, United States, United Kingdom, Germany, France, Netherlands, Sweden, Switzerland, Spain and Singapore.
CrowdMe.One sessions now support file uploads and downloads. The new options are available from the session menu and make it easier to move documents, images and other small files into or out of the remote browser.
To upload a file, select Upload file from the menu. The file is transferred to the remote computer and placed in its Downloads folder. To retrieve a file, select Download file, refresh the file list and choose the file you want to save to your device.
The first time you download a file, your browser may display a Leave page warning. Choose Leave or Yes to continue with the download. The CrowdMe.One session will remain active.
File transfers are currently limited to 10 MB per file, and each session can handle a maximum of one file transfer at a time. These limits are intended to keep transfers reliable during the beta.
Files are transferred between your device and CrowdMe.One using the encrypted HTTPS connection used by the website. During a transfer, a temporary copy may be stored on the CrowdMe.One application server while it is moved between your device and the remote computer.
No additional file-level encryption is applied to this temporary copy. It is deleted as soon as the transfer succeeds or fails. Temporary transfer storage is also cleared when the CrowdMe.One service starts or shuts down.
Files that remain inside the remote computer are destroyed together with the rest of the session when the session ends.
CrowdMe.One now automatically selects a browser resolution based on your device. Sessions started from displays narrower than 1000 pixels use a smaller browser resolution that is better suited for phones and tablets.
The session toolbar now lets you switch between 1920×1080 and 1280×768, and between horizontal and vertical orientation at any time during the session.
A new KB button makes it easy to open the on-screen keyboard on mobile devices whenever text input is needed.
CrowdMe.One can now also be installed as a Progressive Web App (PWA). This adds a CrowdMe.One icon to your phone or tablet and opens the service without the normal browser tabs and address bar, giving more screen space to the remote browser.
Since browser screen resolution and orientation are among the characteristics that may be used for browser fingerprinting, different combinations can produce different browser fingerprints.
One of the easiest ways to understand CrowdMe.One is simply to test it yourself. Open CrowdMe.One from different devices—for example a phone, tablet and desktop computer—and compare the results using the tools below.
FingerprintJS generates a browser fingerprint from commonly available browser information. In our tests, CrowdMe.One produced the same visitor identifier regardless of whether the session was started from an iPad, laptop or desktop computer. This shows that, according to FingerprintJS, different devices present the same browser identity when using CrowdMe.One.
Different privacy tools use different approaches. Some attempt to reduce fingerprinting by varying browser characteristics between sessions. CrowdMe.One instead starts every session from the same browser template, allowing users to verify that consistency for themselves. FingerprintJS is only one fingerprinting method, however, and other websites may use additional techniques.
BrowserLeaks lets you inspect exactly what information your browser exposes, including WebGL, Canvas, fonts and many other browser properties. It is useful for understanding what websites can actually see and why different fingerprinting methods may reach different conclusions.
Cover Your Tracks, developed by the Electronic Frontier Foundation (EFF), estimates how trackable your browser appears compared with other visitors. Rather than producing a single fingerprint, it explains which browser characteristics contribute to browser fingerprinting.
No single fingerprinting tool tells the whole story. FingerprintJS provides an easy way to compare browser identities, while BrowserLeaks and Cover Your Tracks help explain what information websites can observe and why different fingerprinting techniques may produce different results.
Try the tools yourself:
• https://fingerprintjs.github.io/fingerprintjs/
• https://browserleaks.com/
• https://coveryourtracks.eff.org/
CrowdMe.One has been updated to the latest version of noVNC, bringing improved clipboard support between your local computer and the disposable browser.
You can now copy text from the remote browser to your local computer, and also send text from your local computer to the remote browser using the noVNC Clipboard panel.
Open the side menu in the browser window, select Clipboard, paste or type the text you want to send, then use the normal paste command (Ctrl+V or the application's Paste menu) inside the remote browser or desktop.
After feedback that text looked blurry on some displays, the default session resolution has been changed to 1920×1080.
Firefox Sponsored Shortcuts have also been turned off.
CrowdMe.One has entered a beta period.
Free test tokens are being handed out to gather feedback and improve the service before a wider release.