Nice — the use case for this is immediately obvious. I've spent more time than I care to admit running `lsof -i :PORT` or `netstat -tulpn | grep PORT` trying to figure out what's squatting on a port before running a dev server. Having a clean CLI for this with kill support is the kind of small tool that earns a permanent spot in your toolkit. Added to my dotfiles setup.
Ditto. Though I always take the lazy route and just change port numbers until I find an open one. My Mac is probably running like 20 different localhost apps at any given time.
I always have a bunch of local projects running, particularly during the weekend where I'm rarely working on one thing at a time. A big pain of mine was constantly running into port: Redis from one project blocking another, orphaned dev servers from old worktrees, Docker containers I forgot about. The usual fix is lsof -iTCP | grep ..., then figuring out what the PID actually is, then killing it. But I always forget the command, and it doesn’t really include all the information that I like.
So I built this lightweight CLI. Single binary, no dependencies. It shows everything listening on localhost with process names, Docker container info, clickable URLs etc.
Sure there are workarounds, but none that satisfied my need for a short, easily rememberable command. Also nothing really has the same satisfaction as running sonar kill 3000 — it just feels nice. I’ve already been approached by a few agent orchestration tools that have been struggling with the same thing. It's really useful when you have multiple agents running, but it's not built for just that use case, I have also find it handy when killing off all containers after a failed cleanup and so on. Also know that MCPs are dead and CLIs are the new thing in agentic coding, this might be a useful tool for Claude, particularly when a compose process exits before all containers are stopped.
> I’ve already been approached by a few agent orchestration tools that have been struggling with the same thing
Wow, this says more about the agent orchestration tool ecosystem than what you might think, that they're unable to kill child processes they themselves spawn makes it seem like they have zero clue about what they're doing.
Probably why my impression always end up with "Wow, what a vibe-coded mess" when I look through the source of all these harnesses, they don't seem engineered at all.
This is a stupidly annoying problem because it's _very easy_ to accidentally spawn children that won't get killed up in many kill situations because the distinction between processes and process groupes papered over by the fact that shells will be nice enough to kill via process group.
But if your program is some TUI in raw mode its ctrl+c handler is often just killing itself... leaving its children along! Process groups in Unix are a stupid mess and the default being "a process can go away with the subprocess sticking around" rather than the inverse has just caused so many of these long-standing issues.
> that they're unable to [manage and] kill child processes they themselves spawn makes it seem like they have zero clue about what they're doing.
Yeah, at the bare minimum these projects could also use something like portless[1] which literally maps ports to human- (and language model-)readable, named .localhost URLs.
Which _should_ heavily alleviate assignment of processes to projects and vice versa, since at that point, hard-to-remember port numbers completely leave the equation. You could even imagine prefixing them if you've got that much going on for the ultimate "overview", like project1-db.localhost, project1-dev.localhost, etc.
Well, or just use port 0 like we've done for decades, read what port got used, then use that. No more port collisions ever. I thought most people were already aware of that by now, but judging from that project even existing, seems I was wrong.
Can never remember the right lsof incantation for this and end up googling it every time. The Docker container detection is a nice touch — half the time whatever's squatting on port 3000 is some container I forgot to stop two days ago.
Nice! I always forget the lsof flags and end up googling them every time. Would be cool if it could run in the system tray and show what's running on your ports at a glance. Also, the name had me thinking SonarQube at first, might be worth considering a rename to avoid the confusion.
Hey, thanks for sharing this. Your app inspires me to take a look at Go, again! I've been searching for another primary language to learn. My primary used to be Java at $day$ job and now Python for ML/AI. I love Python but still feel insecure given the lack of static typing. I look at TypeScript as well, especially in the context of Bun runtime. I decided it may not be for me, not the language, but the ecosystem around it.
I quite like go for how lean it is, really nice for local projects with lots of constant changes. Also easy to learn. The biggest most annoying thing is how bad error handling is done, not sure why they built it like that.
I primarily write typescript and python for work. But have dabbled in bunch of other languages at different jobs and for different tasks. Yet to pick up rust, but have been wanting to. But tend to pick what seems most right for the task, while also considering what i most want to be working with
I forgot that I also "try" Rust as well. But, I feel like it may not be suitable for my use cases and not simple enough for non-intelligent person like me :-). I agree error handling in Go could be better. But, comparing to Java I don't think I would feel salty.
Would be nice to have a flag to customize the URL displayed for Docker containers. I connect to my host via Tailscale, but I can’t open links with localhost. It would be helpful to have a parameter that allows us to choose a network device or specify an IP address to display.
HN is a place where people can be expected to go beyond the title (though I like the limited script and am glad it was posted). Misleading titles are not uncommonly flagged and changed, even.
True, it was what it started as, but grew as i found my self missing features. Got a few users and now i don't want to update the name. Also easy and quick to write in the terminal
We live in crazy times. I wanted to add a PID to the list for my personal use and since I use Rust way more than Go, I decided to one-shot one app, and Codex indeed one shotted it, wow.
Ahhh yes sure, but not too worried about that i think. Its not so difficult to create, but to innovate and maintain requires participation. It can be the bulk of the job. But most def its sad if people leave their good ideas in a fork of their own. But also great that you can reproduce a good idea thats been conflated with wayy too much crap, because a product lost site of ehat their value proposition was
I'm tempted to one-shot this into a series of FISH abbreviations.
And I would want someone to use that to one-shot a python implementation. And on and on like a game of telephone until the context degrades so far that it becomes an entirely different program.
I always find myself going through my zsh history for `lsof`. Will definitely check this out, seems interesting (even though I'm generally reluctant of installing third-party tools for such jobs).
I get that, i also often install some and forgot about them. But i felt that there was a big gap in managing multiple services running on localhost. It's pretty lightweight if that helps
Can it do things that existing tools build into my distro cannot?
It looks to me like a convenience tool. Which is fine, but i see no need for it myself
Probably not, that's what makes it so lean (it doesn't really pull in anything).
Although i have a feature branch with a tray app (for macos) that let's you monitor and track any process (will send notifications when one goes up or down). But it's just gimmicks i felt i needed to make life a bit easier when working with compose and across worktrees
Christ Almighty I hate our industry practice of binding to some inscrutable port number on localhost. Unix domain sockets aren't that hard! They're secure against all sorts of attacks and more convenient to boot. Instead of connecting to a number, you connect to a file. An ordinary file, with an ordinary name you can mv, chmod, and rm. Boring on a good way.
So why doesn't everyone run local services over Unix sockets?
The only problems: 1) web browsers don't support AF_UNIX URI scheme, and 2) ancient versions of Java don't have built-in APIs for AF_UNIX sockets.
That's it. For these trivial reasons, we've beat our head against arbitrary opaque numbers for decades.
Some random daemon binding to 3000 because it's the express default drives me nuts. I either do a Unix socket, a pick any random port if it has to bind on a port.
Yeah, that too. Windows supports them nowadays too, just to be clear. I think we're still bottlenecked, right now, on #1 and #2 in the form of Java 8 refusing to die.
Yeah, doing the math it's actually only 33 years of not supporting AF_UNIX, but that's not really right either, since those versions of Windows didn't support any sockets. I guess the technically correct answer then is that Windows didn't support UDS for 26 years.
Which is still enough for most portable software to go "eh, localhost is fine*"
* resolving localhost is actually a pretty bad idea (yet very common) and it's way more robust to listen directly on a numeric address.
I think a cli tool that detects objects beneath the surface is a pretty intuitive name, but i was also reluctant in the beginning. But they are pretty keen on always using the qube part, i believe theirs is sonar-qube.
Honestly, pretty cool. I was wondering if something like this existed. Right now I have scripts to kill the ports I use consistently to avoid issues when developing. Kudos!
I looked around for a while, couldn't find anything. Someone posted about killport but never stumbled upon it. Has some features that are the same, but not all. From the reaction online people don't seem to know of other solutions like this, or have something they have made them selves in the .zshrc :)
Yeah, I had to create the script myself and used an extension so I could click a button in VS Action bar. Once I outgrow my current setup, this is probably what I'm migrating to.
Developers are nitpicky, atleast i am and i know a lot of others that are as well. So don't underestimate the value of a nice tool with good developer experience, one that's intuitive, clean and easy to use means a lot when juggling so many things during a workday. So having a clean and light implementation to make job even easier is in my opinion worth it (and thus needed) :)
True, but as i write their are workarounds, the problem is that they are unintuitive, difficult to remember and don't provide all that much usability beyond listing. So these lack useful features like getting process stats, killing ports easily without having to remember the the pid after lsof and so on. I often have to kill multiple process at once after a failed cleanup. If you are into agentic coding, then having your agent create a profile for all the processes it stats, which it can easily kill of when finished is a lot easier for me atleast.
Some features on the way are: next available port; wait (wait for a host to return a successful health check before proceeding - good for migrations etc.). And lots more. It's not just about listing running ports, but a tool for managing them.
But to each their own, that's what's lovely about the many options available. But if you have anything in relation to this you think is neat, feel free to open an issue. It may be able to convince you that a simple alias won't suffice.
Glad to hear! Have quite a few ideas in mind so keep an eye out for some updates (one of the ideas is an easy update command). There's a couple of open enhancement ideas as well. Feel free to add any or contribute.
So I built this lightweight CLI. Single binary, no dependencies. It shows everything listening on localhost with process names, Docker container info, clickable URLs etc.
Sure there are workarounds, but none that satisfied my need for a short, easily rememberable command. Also nothing really has the same satisfaction as running sonar kill 3000 — it just feels nice. I’ve already been approached by a few agent orchestration tools that have been struggling with the same thing. It's really useful when you have multiple agents running, but it's not built for just that use case, I have also find it handy when killing off all containers after a failed cleanup and so on. Also know that MCPs are dead and CLIs are the new thing in agentic coding, this might be a useful tool for Claude, particularly when a compose process exits before all containers are stopped.
Open for contributions, ideas and feedback.
Wow, this says more about the agent orchestration tool ecosystem than what you might think, that they're unable to kill child processes they themselves spawn makes it seem like they have zero clue about what they're doing.
Probably why my impression always end up with "Wow, what a vibe-coded mess" when I look through the source of all these harnesses, they don't seem engineered at all.
But if your program is some TUI in raw mode its ctrl+c handler is often just killing itself... leaving its children along! Process groups in Unix are a stupid mess and the default being "a process can go away with the subprocess sticking around" rather than the inverse has just caused so many of these long-standing issues.
Yeah, at the bare minimum these projects could also use something like portless[1] which literally maps ports to human- (and language model-)readable, named .localhost URLs. Which _should_ heavily alleviate assignment of processes to projects and vice versa, since at that point, hard-to-remember port numbers completely leave the equation. You could even imagine prefixing them if you've got that much going on for the ultimate "overview", like project1-db.localhost, project1-dev.localhost, etc.
[1] https://port1355.dev/
https://github.com/RasKrebs/sonar/issues/15
I have a tiny shell function to kill whatever is listening on that port
I primarily write typescript and python for work. But have dabbled in bunch of other languages at different jobs and for different tasks. Yet to pick up rust, but have been wanting to. But tend to pick what seems most right for the task, while also considering what i most want to be working with
https://github.com/RasKrebs/sonar/issues
I like clack/prompts. See its multiselect API.
https://github.com/bombshell-dev/clack/tree/main/packages/pr...
https://github.com/raskrebs/sonar/issues
https://github.com/fcoury/sonars
sonar list -c port,process,pid,type,url,container
or just show all columns:
sonar list --all-columns
But yeah, it's quite cool. I believe the future lies in software distillation, so cool to see it happen on my own project :D
And I would want someone to use that to one-shot a python implementation. And on and on like a game of telephone until the context degrades so far that it becomes an entirely different program.
https://github.com/clutchski/dotfiles/blob/main/home/bin/por...
For the even less patient there's also this (not mine): https://github.com/jkfran/killport
Started with the same, but found my self wanting a bit more, so just built it
Although i have a feature branch with a tray app (for macos) that let's you monitor and track any process (will send notifications when one goes up or down). But it's just gimmicks i felt i needed to make life a bit easier when working with compose and across worktrees
https://github.com/raskrebs/sonar/issues/15
So why doesn't everyone run local services over Unix sockets?
The only problems: 1) web browsers don't support AF_UNIX URI scheme, and 2) ancient versions of Java don't have built-in APIs for AF_UNIX sockets.
That's it. For these trivial reasons, we've beat our head against arbitrary opaque numbers for decades.
And so, for want of a nail, the Unix was lost.
> The only problems:
3) 40 years of Windows not supporting UDS.
Which is still enough for most portable software to go "eh, localhost is fine*"
* resolving localhost is actually a pretty bad idea (yet very common) and it's way more robust to listen directly on a numeric address.
Some features on the way are: next available port; wait (wait for a host to return a successful health check before proceeding - good for migrations etc.). And lots more. It's not just about listing running ports, but a tool for managing them.
But to each their own, that's what's lovely about the many options available. But if you have anything in relation to this you think is neat, feel free to open an issue. It may be able to convince you that a simple alias won't suffice.