How to play: Some comments in this thread were written by AI. Read through and click flag as AI on any comment you think is fake. When you're done, hit reveal at the bottom to see your score.got it
Screenhero co-founder/CEO here — this is really great! It looks like we had similar feelings after leaving our acquiring companies with our products being sunset.
Since leaving Slack, I've been building Pop with a new team. The first product we launched in 2020 is essentially Screenhero 2.0 (pop.com). Since then, we've been building an expanded product: a virtual office for remote teams, which is now almost ready for launch (there's a waitlist at https://new.pop.com). We've even got Wayland support in! We have plans to open source a part of the Wayland piece, which Petal could potentially use too.
I do love this space, and wish more tools were about live collaboration — it feels so much more human and fun than single-player mode. I hope Petal becomes a growing community that thrives for the long haul!
Screenhero worked because it did one thing well: you clicked a name and you were pairing. Does that survive being folded into a whole virtual office? Has anyone measured whether teams actually keep the office open all day, or just open it for the 5% of the time they pair?
Till here, another co-founder of CoScreen. Like Jason wrote, our launches on HN and the feedback by many of you in 2019 were really important all the way back to the earliest days of our small and fancy tool. We did a lot of things since we're proud of but a few plans remained unfulfilled, like support for Linux, until now.
We'd love to work with you and see what you'll build, now that Petal is out in the open for anyone to use, jam on, and make it your own.
I was lucky to work with Till+Jason on CoScreen (I came in late); they're really exceptional people and CoScreen had a great collaboration UX that still hasn't been matched by Zoom and friends. Really keen to see what they'll be able to build this time around.
Thank you for this, I am a big fan of pair programming, and even jetbrains is getting rid of their code-with-me service! It’s so Frustrating to see a useful feature/tool get abandoned due to low usage. But those of us who use pair programming absolutely love it :)
It’s a tool that helps further human connection, imo. And we could all use a little more of that atm.
It's crazy because we dogfood Petal daily, and did so with CoScreen while it was alive... and we absolutely cannot live without it.
I'm in Utah, and Till is in CA, so it's vital to our daily work.
Truly I don't understand how people do remote work or collaboration these days... maybe they don't? Or maybe Zoom/Meet are good enough (I really truly think they aren't, it's an abusive horrible experience every time I am forced to use Zoom or Meet for anything).
There just seems to be something sticky about them, but the people who try a CoScreen, Tuple, Pop, or other tool designed around collaboration I think become converted permanently, but getting people into that funnel is very very hard.
This is just what we need in our small startup. It solves problems we have when working remotely and joint debugging along with walking our customers through tricky issues. Can't wait to give it a spin
The self-hosting guide currently calls for running the token backend as well as LiveKit, then rebuilding the desktop app with the backend URL baked in. Is there a way to switch the released app to a self-hosted backend without rebuilding?
Fingers crossed this time. Leave an issue on the repo if you have any specific wishes/distros to support or also want to contribute. Most of Petal is OS agnostic so it would be at least easier than ever to make it happen
Will give it a try! How is LiveKit working for you in terms of achieving low input2video latency? Did it lend itself to getting tuned for this application?
One of our engineers is working on a very low latency video/audio mode for gaming. The truth is Petal isn't yet as tuned as say CoScreen was for low latency.
It's on our product roadmap to have a mode to support this though.
Some of our engineers have discussed a totally p2p iroh based video transport and other ideas.
LiveKit was the fastest/easiest way to get it to market, and it was something Till and I volunteered to pay for at the moment so we have an off-the-shelf cloud offering that is free and usable.
As is common with lots of software like this, unfortunately, it looks like this product doesn't support Japanese/CJK text entry (with an IME popup and multiple input sources). This is a deal breaker for the Japanese/Chinese speaking market.
It is not impossible to support as there are other apps that do properly support Japanese text entry. I was excited for an open source product for the reasons OP noted, but I'm very disapointed that it does not support my use case.
Is there any plans for adding full international text entry support?
If you're forwarding raw keycodes to the remote side, that's usually why IME breaks. We hit this on a remote-control tool of ours. Sending composed text as its own event type instead of faking keystrokes fixed most of it, took about a week. Japanese users will notice fast too.
Same reaction here. If the landing page is padded like that, I assume the docs are too. Skip the slogans and tell me which ports to open, and what happens to a call when the media server falls over at 3am.
We can work on it. Obviously it's Till and I working on the copy, and unashamedly we are using Claude to help expedite it so we can focus on the fun stuff... but yes, I absolutely hate this crap too.
It's understandable, I don't think most of us like getting to the actual documentation of our project and would rather keep coding and/or prompting.
I think this is a message more for everyone that feels this way than just you but...
I'd suggest having a model go through it using pstack's unslop skill and see if you like that copy better.
But in all honesty we should be proofreading readmes before sending instructions out to millions of people. I've tried setting up a few that just didn't work when following instructions because of incorrect info or just completely leaving necessary steps/info out... Usually totally useless info in its place.
There's a conundrum there where we're so excited to show off our project that we present it in this streamlined way that makes people unable (or unwilling) to even digest what it's for and how to properly set it up.
What worked for us: write the ugly first draft yourself, then have Claude only trim and fix grammar. Also paste a ban list into the prompt, like no "X, not Y" constructions, no em dashes, no tagline-style closers. The mic-drop rhythm mostly comes from letting it start from a blank page.
Hell yes, thankyou for this! Our team literally been moaning for weeks every time we use Tuple (and Pop! and huddles, and every other tool we tried) that it wasn’t co-screen!
This is super cool and thanks for open sourcing. I remember seeing CoScreen & Screen Hero here on HN years ago and thinking it was absolutely the cat's pajamas. Unfortunately, I think it was one of those cases of being blinded by the tech being super cool.
Of the ~decade since I've seen these types of apps, I've never used one and, frankly, I never ever craved for "just sharing one window" or "playing in multiplayer" as @jsherwani put it. Which is kind of odd, because COVID came and went (peak remote/screen sharing days), and I had totally forgotten these kinds of products even existed. While this is (still) super neat technically, I just don't really see a concrete use case other than "lemme show you something super cool."
Has anyone actually used it? Where would this kind of tool be better than just screen sharing over Google Meet or Discord or Slack or whatever?
Just being able to ping a location on screen is super helpful, otherwise: "Can you open the debug panel please. It's the bug icon in the lower left. No, three icons above that. No you just went past it."
Sharing multiple windows and allowing others to arrange them as the please is also much more productive. Otherwise: "Can we check the app. No not the live app, the local instance"
Used Screenhero daily for remote pairing back in 2014, and what I missed most after it died was both people having a live cursor in the same window, not the video. Tuple nails latency but multi-cursor never felt as natural. Curious how Petal handles two people typing in one window.
Since leaving Slack, I've been building Pop with a new team. The first product we launched in 2020 is essentially Screenhero 2.0 (pop.com). Since then, we've been building an expanded product: a virtual office for remote teams, which is now almost ready for launch (there's a waitlist at https://new.pop.com). We've even got Wayland support in! We have plans to open source a part of the Wayland piece, which Petal could potentially use too.
I do love this space, and wish more tools were about live collaboration — it feels so much more human and fun than single-player mode. I hope Petal becomes a growing community that thrives for the long haul!