• Album@lemmy.ca
    6·
    3 days ago

    It’s really hard to not have a unique fingerprint in a casual browsing experience. Even with minimising techniques because some of that will make you stand out and so the easiest way is to look like the crowd. You only need a few pieces of low entropy information to make you stand out. Spoofing is detectable and so can be discarded when generating the fingerprint.

    Tor browser plus ubo is going to be your least fingerprintable route.

    After that I’d say Firefox with arkenfox’s user.js plus ubo.

    You can get Fennec + arkenfox on Android which is Mull Browser.

    Arkenfox uses the RFP feature you enabled plus some other things like letterboxing so your monitor resolution doesn’t give you away.

    If you’re not using tor you need VPN or your IP is going to give you away.

    I suggest reading the arkenfox wiki for more info.

      • boredsquirrel@slrpnk.net
        1·
        2 days ago

        You want to

        1. Reduce the amount of data sent
        2. Make the data less unique
        3. Scramble often scraped data to resemble current browsers, GPUs etc

        There are addons called “fingerprint defender” which scramble font, gpu details and more. I think they are good but have not looked at the code.

        The danger with minimized data is sometimes that the data may be outdated or trackable.

      • MrOtherGuy@lemmy.world
        1·
        2 days ago

        That’s not necessarily a good solution either, because a service could figure out that the source of random fingerprint data likely comes from the same user. Especially if your ip is not changing. It might perhaps be effective if a substantial amount of people were doing it though.

        But to generate such random fingerprint is difficult because it consists of many parts and services don’t all build fingerprints the aame way. You could easily randomize e.g. canvas data, but the issue is that if you only randomize one data point then that one random data point pretty uniquely identifies you if your other datapoints are stable. So to be effective you would really need to randomize several different datapoints and that may not be such an easy task since websites could build them in all sorts of ways.