What is your PGP setup?

Speaking to a few people on matrix, I have found that people have very different, but very innovative setups for PGP. So, I thought, what do the other people of the privacytools forum have as a setup?

1 Like

Till very recently pgp with brew. Now gpgtools w/o the mail part. A lot easier now than terminal-only.

edit: just added one word

1 Like

I use it with Seahorse. https://www.openpgp.org/software/seahorse/

I use gpg4win as my main PGP software and testing PGP Desktop.


I’ve generated a PGP key on the Linux terminal, but that’s about it. Can you perform all of the same functions as a program like Seahorse or GPG? (I imagine you can; I just haven’t done it yet.) To be honest, I’m still learning the terminal and haven’t mastered it yet.

I didn’t like any of the GUI’s for key management, and some had limitations (ie. no 4096bit keys) so I resorted to the command line for creating and managing keys.
Other than that, mailvelope is pretty cool for webmail when I’m not using protonmail.
It’s super easy to use.

Sorry forgot the link https://www.mailvelope.com/en/

Don’t let the lame website fool you, it’s pretty neat.

1 Like

I apologize for thread necromancy but I was asked to post in this by request.

My PGP setup consists of a Talos II OpenPOWER Workstation and a “Redzone” made from an old Lenovo T400 laptop equipped with a Chaoskey and with a patched kernel that had its options set to distrust the CPU RNG. Depending on which secret keys I am using I may use between four or five separate Yubikeys, including one for my root certifying key, one for my personal certifying key and another for its subkey, and possibly more for other purposes or short-lived keys. The Redzone is only used by request only and has had its microphone, speakers, camera, bluetooth, and networking cards physically removed and glued shut. It is only powered on when running on battery power.

To ferry data between the “Redzone” and “Blackzone” (area where plaintexts are handled and viewed, and area where ciphertexts are handled, respectively), I would burn CDs and physically put them through the shredder when they are no longer needed. The Redzone is trusted only to view plaintexts and encrypt with session keys, but is not trusted to generate long term secret keys. I chose CDs rather than USB drives to mitigate several attacks that use USB drives as a vector for malware or devices such as microcontrollers that can conduct DMA or device spoofing attacks on the computer they are attached to: the Redzone has had its USB slots plugged with glue and sealed over with tamper-evident red sheathing tape for this reason and only the laptop keyboard is used.

My keys are generated on the Talos II. Since there are no LiveCD distributions for POWER9 at this moment, I will install a fresh copy of Debian (this was chosen for their emphasis on reproducibility) with FDE and with the network and BMC unplugged on an old spinning disk hard drive and then after provisioning the Yubikeys and smartcards, the LUKS header is written over and the drive wiped, then I shred the 12-word password.

This isn’t a part of the setup itself, but when using the Redzone, I experimented with GPG ratcheting, by certifying a set of short-lived keys with my root key, then each message changing the encryption subkey in order to achieve forward security, then revoking and leaking the signing secret key after the exchange is done. This only works though if both people agree to do it or if my sender has only encrypted it for me without encrypting it for themselves as the default recipient, so while I’ve drilled and practiced with it, I haven’t actually done it to follow a specific security need. I am at this point looking for a way to inspect other state-carrying elements on the Talos II outside of the hard drive.

This setup is somewhat experimental and I did it while I was attempting to document the process of setting up an airgapped viewing station for people with much higher threat models than me. I feel like I learned quite a bit over the course of planning out and assembling the setup, and while it was educational, I don’t feel that it’s practical except in a number of very specific situations. I’ve also been considering ways to improve it and recognized in some ways I could have done better, but this is about as far as I will go for the sake of documentation.

It may strike you as slightly ironic after all this to find out that I prefer to use Signal, but that’s another talk for another day.


I tried to use PGP but found it difficult to manage. Managing the keys and ensuring the authenticity of a message through signature is good, but the process seemed weak to me. I’m happy with the way Tutanota does to create a encrypted message channel to non-Tuta users. Right now I just use PGP to verify software integrity.

1 Like

I just started using Mailvelope! Very helpful. Do you normally use the terminal for PGP?

Someone else had told me that they thought PGP was outdated as an encryption system - yet quite a few people still use it! I’ve started suggesting using messengers like Signal or Briar instead.

1 Like

@virticalizes yes terminal most of the time and then mailvelope for email sometimes :slight_smile:

1 Like