Ghostling | Hacker News Reader

Ghostling

github.com

333 pointsbjornroberg70 comments

I use libghostty for Trolley[0], which packages TUIs as desktop apps, like Electron does for web apps.

It really is quite an amazing piece of software. I just wrapped it in a useful GUI and a bundle/package CLI and it just works. Even on Windows. Kudos to the Ghostty developers.

[0] https://github.com/weedonandscott/trolley

I think your github readme is really missing a picture/screenshot to quickly understand what is the experience like. I.e. if your app is mainly about adding the chrome (as in the surrounding UI pixels) around the TUI, then it would be good to show what is the chrome like.

Nah, I think it’s pretty clear. It would look like a terminal emulator. Just like how Electron looks like a bunch of browser widgets - because it’s literally a single-web-app browser.

Does it have any extra controls? And is it using native windows (as opposed to Electron)?

A screenshot would still make it clear that that's all you're getting - and no extra chrome etc

I agree a screenshot or even better a video of lazygit, btop or something similar running in the tui side by side with the trolley app

This is a pretty cool idea. Kind of a neat distribution hack if all you have is a TUI (and not a full GUI). Curious whether you know of any success stories yet

Thanks.

This was written for my own screenwriting software, which is now in private alpha. It works quite nicely for an alpha

https://blisswriter.app

I kind of want to use this to turn Wordgrinder into a Mac app haha

Oh this is nice!

You mention that Android/IOS support is possible. I hope that you can please add these support properly as your project matures.

Asking this because I would love to have a cli tool where I can just point to for example golang codebase from any device (mac/windows/linux) etc. and thanks to golang's cross compilation simplicity, have it be compiled for android (well linux arm fwiw) and then have it all be compiled into a single android apk.

And if you do that, I would love to have something like zenity but for android so as to abstract the cli behind a nice gui for mass-adoption.

This is almost a million dollar problem as there are so many good cli tools and its incredibly easy and versatile to make a gui even with scripts on top of that cli but Android/Ios usually don't have that versatility.

The C file is small enough to read (over a few minutes.)

I got to about line 5 and realized: I’ve never seen quite that technique for embedding a font via an autogenerated header before. I’m more used to Windows resources; this seems to generate a byte array in CMake code. I’m somewhere between horrified and impressed, in that I feel we’ve finally discovered a cross platform binary resource embedding solution.

And as a Windows programmer the use of a method called DrawTextEx surprised me :)

A really neat sample. Shows the power of the ghosttty library very well. The author chose well with their other libraries, it’s the kind of demo that lets the code actually demo what to trying to without much else getting in the way. Rather inspirational to wrote my own terminal app now.

Fun fact: XPM bitmaps were designed to be #included unmodified, the files contain C boilerplate: https://en.wikipedia.org/wiki/X_PixMap

Here's the build script that uses: https://github.com/ghostty-org/ghostling/blob/main/bin2heade...

I ran it against a 1x1 pixel GIF:

  cmake -DINPUT=pixel.gif -DOUTPUT=pixel.h -DARRAY_NAME=pixel_gif -P bin2header.cmake

And got this:

  // Auto-generated from /private/tmp/exp/pixel.gif — do not edit.
  static const unsigned char pixel_gif[] = {
      0x47, 0x49, 0x46, 0x38, 0x39, 0x61, 0x01, 0x00, 0x01, 0x00, 0x80, 0x00, 
      0x00, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x00, 
      0x01, 0x00, 0x01, 0x00, 0x00, 0x02, 0x02, 0x44, 0x01, 0x00, 0x3b
  };

Interestingly, cmake 4.3 just added a builtin command 'bin2c':

https://cmake.org/cmake/help/v4.3/manual/cmake.1.html#cmdopt...

...it will probably take a decade or two until Debian-based Linux distros will get cmake 4.3 though ;)

For cross-compilation ease it makes sense if you don't care about the size explosion.

You can use `xxd` from the vim package to generate these. You'll find out pretty quickly that this is only suitable for small resources: gcc and clang blow up in both time and space on very large literal arrays. If you need to ship more than a few megabytes, find a different technique.

I used this technique for awhile, but it was too problematic for my use case. Now, I use https://github.com/lief-project/LIEF -- among other things, this project can modify Windows PE, macOS Mach-O, and Linux ELF binaries to add resources to them, then offers an API to read them back later. It's a little different for each format, but it's capable of doing all three and I was able to build a cross-platform resource-bundling system that doesn't care how big the resources are.

Yeah xxd is the correct answer. If you don't want to install a dependency and have Lua installed, or if you're just feeling a little bit frisky, you can use my function which is Production Ready™.

    Xxd = function(name, input)
        if not name:find'^[_%a][_%w]*$' then error('bad name: '..tostring(name)) end
        local ans = {
            'const unsigned int '..name..'_len = '..(#input)..';',
            'const unsigned char '..name..'[] = {',
        }
        local t = {}
        for i=1,#input do
            table.insert(t, ('0x%02x,'):format(input:byte(i)))
            if #t == 16 then -- 16 columns per row. arbitrary, change this if you want
                table.insert(ans, table.concat(t))
                t = {}
            end
        end
        if #t ~= 0 then
            table.insert(ans, table.concat(t))
        end
        table.insert(ans, '};\n')
        return table.concat(ans, '\n')
    end

I am distributing it under the terms of the GNU GPL v3. So if you put this in your codebase I will sue you into releasing your entire source. Just kidding it's MIT licensed.

Honestly that's a terrible joke. Seriously it's MIT. Here I will put the full license in this comment to illustrate how serious I am:

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Feels like Lua is a more exotic dependency. I used to use xxd but this gets problematic when files grow (and they don't need to grow much at all), objcopy is much faster (which I didn't think would matter much, but it did) and don't have the same issues that accidentally opening the .h file in your editor or code-searching having to traverse that mess.

Yes, you'd want to gitignore it and exclude it from search etc. but you still have size issues etc. See gucci-on-fleek's comment on objcopy above for usage.

For some things there is nothing better than xxd. It is so simple.

> I’m somewhere between horrified and impressed, in that I feel we’ve finally discovered a cross platform binary resource embedding solution.

Maybe I'm misunderstanding your comment about having "finally discovered" that...

For to me embedding binary resources in source files is nothing new at all? It was definitely done in the early JavaScript days for a variety of reasons.

Arguably early basic listings that had lots of DATA text lines were already doing that. Maybe not the most portable but we're talking about the 70s and 80s here and definitely binary data in source code.

Games for the Atari ST and Amiga, for example, could partially share at least some of their source code and it wasn't uncommon to encode binary inside sources, including... Fonts! Back then fonts were not reused from one game to another.

Heck, I've done in Java (and Java is cross platform) in the past: quick visual debug tools for Java GUI apps to toggle a debug mode showing something not dissimilar to a HUD. Pixel-perfect fonts encoded in .java source files.

I really don't think it's anything new.

P.S: I'm pretty sure it's done like for some fonts in the Linux kernel too.

`man xxd`

"Finally discovered" was more that there are so many difference ways, all of which work differently. Maybe this is just adding one more technique, xkcd-style, but it struck me as more possibly cross-platform-identical than several techniques I've seen. If you embed Windows resources, for example, that's highly Windows-specific; this one would work regardless.

Well, I originally used C23's #embed directive (https://en.cppreference.com/w/c/preprocessor/embed) but GCC in Nixpkgs doesn't support C23 (or I'm holding it wrong) so I dropped back to this. The better long term solution is #embed.

I've definitely used C23 via Nixpkgs long ago at this point, did you use the `gcc` package and `-std=c23`? Both unstable and 25.11 should support it on all currently packaged GCC versions.

I think this has been around for a while:

  $ echo 'test' | xxd -i -n foo
  unsigned char foo[] = {
    0x74, 0x65, 0x73, 0x74, 0x0a
  };
  unsigned int foo_len = 5;

(edit: 30 years)

For completeness, there's also the somewhat common way of creating an object file directly from the binary using objcopy. The object file ends up with 3 symbols with names based on the input file. For instance, a binary file level0.map yields: _binary_level0_map_start, _binary_level0_map_end, _binary_level0_map_size.

These can be used in the application through external declarations; and you include the object file in the linking step, like any other object file.

My preferred technique (where #embed isn't available) is to use objcopy:

  $ echo 'Hello, world!' > hello-world.txt 
  $ objcopy --input-target=binary --output-target=elf64-x86-64 \
      hello-world.txt hello-world.o
  $ cat <<EOF > embed.c
      #include <stdint.h>
      #include <stdio.h>
  
      #define EMBED(NAME)                                \
          extern const uint8_t _binary_##NAME##_start[]; \
          extern const uint8_t _binary_##NAME##_size[];
      #define DATA(NAME) _binary_##NAME##_start
      #define SIZE(NAME) (size_t)_binary_##NAME##_size
  
      int main() {
          EMBED(hello_world_txt);
  
          printf("%.*s", (int)SIZE(hello_world_txt), DATA(hello_world_txt));
      }
  EOF
  $ gcc hello-world.o embed.c -o hello-world
  $ ./hello-world
  Hello, world!

I used ghostty for the shell in my agent-manager tool (think something like https://air.dev/ but in SwiftUI. One architectural detail i'm still going back and forth on is: who should own the PTY?

If embedded Ghostty owns it, now i have to have some kind of workaround to instrument what's happening inside the terminal - signals, stdout/stderr events, return code etc.

If my parent agent-manager app owns it, now i don't have the nice clean small interface to ghostty (i'm a fan of John Ousterhout style narrow but deep interfaces, pulling complexity down rather than pushing up to parent).

Not sure if any other ghostty embedders might have advice. It's on my todo list to task a an agent to "gh repo clone" a few ghostty using shells and report on their arch.

Sounds like you need to multiplex a terminal with a terminal multiplexer.

terminal multiplexer

term mux

tmux

I switched to Ghostty a few months ago and it's become one of the apps I never close. The rendering speed is noticeably better than iTerm2, especially with large log outputs. Excited to see libghostty enabling projects like this — the idea of packaging TUIs as native desktop apps is really compelling for indie developers.

I built my own macOS terminal app over the last two weeks using swift/appkit + ghostty's zig library.

I basically just wanted vertical tabs and notifications for when AI agents (claude code, codex) are finished.

I already use it as my main terminal over iTerm.

It's a fun project since I use my terminal all day, so I always have ideas for how something could be improved or polished. AI can do the chore work of figuring out how to impl some bugfix or UX polish, and I manage the last 10%.

This would have been too much work to maintain for fun before LLMs.

FWIW most of the agentic CLI tools also let you setup notification hooks so you can just do something like this:

  afplay /System/Library/Sounds/Glass.aiff                                                    
  osascript -e 'display notification "Waiting..." with title "Coding Agent"

The sound is especially useful since for longer running tasks I'll often do some chores around the house while waiting for it to finish up.

My notification hook pipes an OSC 777 (title, body) message into /dev/tty, Ghostty handles/ingests it and then emits the event for my terminal app to do things like craft a macOS notif and style the tab/pane of the originating terminal.

    #!/usr/bin/env bash
    MSG=$(cat | jq -r '.last_assistant_message // empty' | head -c 200)
    printf '\e]777;notify;Claude Code;%s\a' "${MSG:-Claude finished responding}" > /dev/tty

I tried the osascript solution first but had some issues, iirc no good way to focus the originating terminal pane on notif click.

Something I never figured out is why Claude Code's "Notification" hook waits minutes to fire. I had to use the "Stop" hook for actual end-of-response timing.

> iirc no way to focus the originating terminal pane on notif click.

Yeah mine has the same issue - it just brings up the script editor. It's not as much of an issue for me since I'm rarely running more than a single Opencode instance at a time.

Could definitely see that being useful if you were running quite a few agents to let you hone in on the correct terminal window/tab quickly though.

Yeah, exactly. I have so many agents running these days. 1-2 per project.

Often one is polishing a plan file for the next feature/bugfix (claude), and the other is reviewing it to give feedback (codex). And that is my life now. But on the upside, it's trivial to maintain sideprojects now.

Nice. You have a similar setup to mine. I've found OpenAI's gpt-5.3-codex (at xhigh) to be pretty dismal at actual implementation but it's surprisingly adept at performing a high level feature branch review and catching issues that are missed by Gemini/Claude.

Yeah, through trial and error I found that Codex (now gpt-5.4) has really good feedback once there's a plan file.

Kinda like someone on the internet who suddenly becomes an expert once you make concrete claims for them to shred into.

I have an AGENTS.md prompt that specifies how to review a plan. Something like: for every finding, rank the top solutions internally and then recommend a solution. And if there's a simpler, better pivot that the plan should take, pitch it at the end.

It's really actionable, and I basically can paste the review back into Claude, then tell Codex to re-review the plan, and repeat this until it has nothing more to add, and now I have a really good plan.

Nice. I have a similar loop (Opus 4.6 plan <-> Codex xhigh review) that I kind of run in an iterative feedback loop until meaningful actionables converge down to zero.

I initially gave some thought to automating this because CC/Codex do have SDKs... but every once in a while Codex will propose some absurdly over-engineered stack advice that I have to manually reject before passing it back to Claude.

Man, we are on the the exact same page.

Yeah, I also thought about automating it before. Especially when I'm middlemanning noncritical plans where I don't care what the solution is.

I did make a skill /ask-codex that gets Claude to call Codex for one round of feedback which is pretty useful.

But for critical plans, I suppose being the middleman gives me a good opportunity to learn about the solution and its trade-offs as it's being incrementally refined. Especially in a domain I don't know much about.

I hardly think you can say YOU built it went all your did was manage 10% of the project. An LLM built it, you just test it.

Figuring out what to build, how it should work at a high level, and what it should be like to use it is the hard part and the defining aspect of any tool or application. (Or a fun game, for example)

Now that the cost of writing code is ~$0, I think the entity who figures that out is the one who built it, not the entity that does the trivial chore once the hard decisions have already been made.

I really hope libghostty succeeds at setting a new baseline for what terminal emulation should be. Seems to be on a right track.

This is amazing and might be exactly what I’m looking for my own weirdo retro tooling that sometimes needs to run over ssh but also expect a “GUI” experience… any metrics on the overhead this might add to, for instance to a hello world type program?

I have an idea of a terminal emulator where you could maximize panes but using a nested structure, does anyone know of one?

Standard "Zoom" features in tmux or iTerm2 only maximize the single active pane to the full window, hiding everything else. If I have a layout like this:

  _____________________
  |         |    B    |
  |    A    |---------|
  |         |    C    |
  |_________|_________|

And I expand B, I want A to hide, while B and C remain visible together. Then I can create a new nested workspace in there and later zoom out when I’m done.

Maybe this could be done arbitrarily deep?

It's not really what you want but in a similar vein check out: zoom.el

https://github.com/cyrus-and/zoom

I guess it's also the kind of thing where screen real-estate is a must otherwise when you're at the "top-level" it would look weird-ish.

You can do this with i3

Wouldn't you want a hide option rather than a fullscreen option? so hide A?

When I use zellij, i just move A to a new tab temporarily

This should be simple any time you have panes in a binary tree split view.

The hard part is the UX: making it clear that we're in a zoom state (esp with nested zooms), somehow showing the split-pane tree so that user knows what container will be filled when they zoom, etc.

Fork ghostty and get claude to one-shot it so you can see if you like the idea.

This looks interesting.

I don't need my terminal emulator to support tabs, windows, or session management. My WM manages tabs and windows, and I use tmux for sessions, which also gives me a scrollback buffer, selection, clipboard, search, etc. This combination allows me to use any simple terminal emulator, such as urxvt, st, and now foot, without issues.

Ghostty didn't appeal to me, but I might give this a try. It's good that OSC support is planned. A plugin-like system, similar to st's but less cumbersome, would be nice to have.

It's comical how much time I've spent convincing people that tabs are a window manager feature not an application feature. People in the Alacritty issue on the subject were pissed!

Are there any good, non-tiling window managers that support tabs? (I struggle with tiling ones like i3 because I am a small-brained mouse user)

I mean, macOS supports tabs now. I wouldn't call it "good" though.

Maybe you'd like Niri?

https://github.com/niri-wm/niri

Fluxbox has tabs and is a stacking window manager.

- https://fluxbox.org/features/

Yes, we need tabs for RDR2 and Spotify.

I would love tabs for Spotify. I just discovered I can at least open new windows from the linux YouTube music client by middle clicking, a revelation !

Every application (or concept) can introduce “tabs”, but it means something wildly different for that particular application. Tabs (or instances) in an application immediately bumps into the concept of state (statefull vs stateless) in applications.

Sometimes, it makes perfect sense. The reason tabs made sense for web browsers since 2004 is because each web page could be thought of as a “stateless” instance of an application. You’re not asking for “tabs”, you wish every application could be “Stateless”. Stateless is a beautiful thing, until you understand what state is, and who needs to manage it.

If every “tab” of Spotify had no idea what the other “tab” is playing and you had to switch back and forth between tabs to pause-and-play songs, that would be a bug, not a feature. While 2 “windows” playing audio (if you instruct them to) is expected.

> Tabs (or instances) in an application immediately bumps into the concept of state (statefull vs stateless) in applications.

Agreed, and this is why tabs need to work at the app level, not window manager/os level.

That said, for Spotify specifically, it can already tell what I'm playing on an entirely different device. I think they can handle tabs.

I've heard this a lot on HN over the years but it doesn't make much sense to me. Some thoughts:

1. App tabs improves UX for 99.999% of users who aren't using a WM with a good tab solution (if one even exists).

2. WM tabs means launching a new app instance for every tab you might want vs having lightweight app tabs.

3. App tabs can do all sorts of app-level things and UX polish that dumb WM tabs can't do because they are so general. My terminal emulator tabs show a badge count of bell notifications, can be dragged around into groups, or dragging into other tabs as split panes. My browser tabs show you which tab is playing music and can impl right click -> mute.

4. I bet even the biggest WM tab cheerleader still uses browser tabs.

5. WM tabs are a different concern than app tabs, not a replacement. WM tabs are useful when you want tabs and the app doesn't provide a good tab metaphor or when you want to tile/group app instances a certain way. That doesn't mean it's not useful for the app instances themselves to have app tabs when it makes sense.

Agree on all the points, except 4. There are even people out there who use lynx as their primary browser :)

Although while I usually like tabs for most apps, I don't use tabs for terminal and rely either on window manager or tmux. I guess the difference is that I often want a mix of tabs and having multiple terminals side by side, whereas I don't really need that for a browser (or very seldom)

Which window manager do you use?

Sway had the better, though often tedious, WM tab solution that I've tried. Niri had a useless one.

I really tried to love sway splits and tabs for terminal windows. But I finally admitted I'd rather just alt-tab to a few different terminal apps, each with its own concern (maybe one per project, this one for my remote machine), and best of all, each with its own internal tabs.

That said, tabs in kitty and tmux, for example, are so basic that you don't necessarily lose much if you were to use WM tabs instead.

On the other hand, tabs in iTerm2, Ghostty, Cmux, probably macOS Terminal -- a bit more powerful and intuitive since you can do things like drag them, and they can show info like terminal state. And in some of those apps, they can be displayed vertically which is my favorite.

My favorite is still AwesomeWM, although I haven't used it in years. Nowadays I'm on mac and my "tiling manager" is about 6-7 custom functions I wrote in hammerspoon. Basically just tile left/right, full size, "reasonable" size + centered, and a shortcut that distributes all my apps in the virtual spaces I've assigned to them.

One of the reasons why I use tmux rather than i.e. iterm2 tabs is that I don't have to change any of my habits or learn extra shortcuts when I'm working with ssh.

I like to manage tabs and windows through tmux and it suits my workflow very well. What are you going to do now?

On tiling WMs I use rxvt-unicode with no window decorations, no gaps, 1 px border, no scrollbar. Then tmux does the rest, namely tabs and splits. Automatic session saving has been a life saver on more than one occasion.

Interesting you mention tmux because it itself resembles a terminal emulator. It has its own terminal feature matrix that controls what your parent emulator can render. It sounds like you aren’t using tabs and splits in tmux but it does include them.

It sounds like you could get away with using a tool like https://zmx.sh which only handles session persistence (attach/detach). It also uses libghostty but only for state restoration on reattach.

A poor man's version of session persistence can even be managed with tmux plugins like tmux-resurrect and tmux-continuum.

I’m seriously interested in this. I wonder if i can use this along some decent gui library and an llm to vibe-code a SecureCRT replacement.

SecureCRT is awesome but it’s crazy expensive :(

I love seeing an `AGENTS.md` in open source projects.

It's now my #1 heuristic to know if the team is on the right track.

(I need to start adding them to all my projects.)

I have plans to add it to my game (SDL).

Reply on news.ycombinator.com