Unless you are space-constrained. Then something like RaspberryPi Compute Module is hard to replace.

It just means a leader or a guide in latin IIRC

Do you know what tech stack do they use? The website only mentions custom Direct3D renderer.

>Do you know what tech stack do they use?

If you are asking this kind of question you have already lost the plot. You don't need a "tech stack".

Not OP, but I've moved back to kitty because of ghostty horrendous startup time on linux

That's a one time cost though isn't it? It's a tradeoff, but once it is booted it's blazing though.

When I use the terminal, I spawn and kill a lot of instances for quick tasks. Having a longer startup times is definitely noticeable.

My login session starts a background ghostty instance, then each new window is almost instant when I press my i3 shortcut.

This process is documented with systemd, but it's not a requirement, a plain xsessionrc could work similarly: https://ghostty.org/docs/linux/systemd

> I spawn and kill a lot of instances for quick tasks.

Why? Do you open a new browser window when you want to visit a website?

I do whenever I want two websites displayed next to each other or when I want to have them in separate workspaces, which is pretty often.

Yes actually. Ctrl+Alt+b opens a new browser window quickly and I go somewhere. I have it faster in muscle memory than switching to browser and ctrl+t and then opening whatever.

Same with terminal. My terminals (xfce4-terminal or kitty -1) both open so instantly that I can ctrl+alt+t and even while typing at high speeds, the next character is entered into the terminal.

I can type the below sequence of keys at high speed (realistic example that OCRs clipboard image):

  ctrl+alt+t vpng | ocr | ctxt alt+tab

without risking the startup time swallowing my `vp` keys

And yes, I disable animations for any kind of window switching mechanism including alt-tab/worksapce switch

At any rate, I believe ghostty has the same single instance mode equivalent.

In case you're not aware, there exists:

    ghostty +new-window

tmux?

I figured that was an issue with the appimage for me. Did you install it a different way that you ran into that issue?

also slow startup times, archlinux, official repo package

I'd say 100-200 ms is not horrendous but it can be much better without giving up on libghostty-vt :) check out Monstar.

Now, Monstar has a bizarre issue (at least in my setup) where it displays blank instead of capital A when used with Fixedsys Excelsior, for some reason. But hey it starts in under 10ms.

edit:typo

I use this icon too! Only thing I customize after installing kitty.

I'd love to use Ghostty, but the start-up time on Linux is unacceptable. I know that's because of GTK, so I hope that maybe Rex[0] woudn't have this issue.

[0] https://www.superlogical.com/

The only think I expect from a terminal is to be fast!

I think there’s a systemd service you can install, a user one

Their model is to have a persistent server running.

Can you elaborate on that?

I'd love to NOT expose my services to the internet, but I often share stuff from opencloud/seafile or immich. If somebody has a convinient (for non-technical people I share stuff with) solution, please share because as the author of the post said, with AI assisted hacking now I'm really concerned. I hope that dockploy is secured poperly.

Shameless plug...

MatterMost was actually a great file sharing platform for our business. But they was not true open-source and removed features we needed like SSO.

So instead we vibe-coded a MatterMost Slack alternative which we have been running on the internet fine for months now https://intellistream-datahub.github.io/intellistream-chat/ Its super convenient to share files on. And it comes with built-in secret sharing like YoPass. We occasionally implement new features and fix bugs we discover.

Its one jar file, and supports Vault/OpenBao and requires Keycloak and PostgreSQL. Lucene is embedded.

Their previous best model - Gemini 3.8 Flash (also when set to high) was even faster. A appreciate it that they focus on speed. I's great for LLM pair-programming.

It's nice to have it available for more complex tasks, but I'll stick with 3.8 Flash thanks to its high output speed. I really like it for fast iteration and prototyping.