#🖥-vesktop-development
1 messages · Page 37 of 1
🤷♂️
kaboom the pr then
i mean i can do
[class*="submenu"] [class*="scroller"] {
this is bugged because it's not defined
just define custom-floating-layer-max-height to fix
thats fixes it yeah
#textarea-context-vcd-spellcheck-settings {
--custom-floating-layer-max-height: 50vh;
}
actually
#textarea-context{
--custom-floating-layer-max-height: 50vh;
}
done
it makes little sense to fix it in vesktop tho
vesktop is slow moving
this is just a discord bug that also affects vencord
we should either just wait for discord to fix it or if you really want to fix it, it should be fixed in vencord because vencord is also affected by it
this for example
and this
should be moved to vencord then, is there even a long enough dropdown in regular discord for them to actually fix it
yes
like this
clearly they are working on this because it was working a few days ago
they borked it with one of their changes
oh i thought this was broken for a while
might aswell wait to see if they fix it then
well i dont mind fixing it ourselves
but not in vesktop xD
vesktop wont get a new release for another month at least
do you want me to pr vencord 
we already fix a similar issue in vencord for BetterSettings but now they broke it even more
fullHeightContext.css: Lines 1-10
/*
* Discord has dumb max height logic for their context menus.
* If a context menu is at the bottom of the screen, its submenus are capped to its max height and can't even grow upwards
* We unset the variable they use to cap height. This allows submenus to grow as tall as they want
*/
#user-settings-cog,
[aria-activedescendant^="user-settings-cog"] {
--reference-position-layer-max-height: initial !important;
}
discord things 
idrk where i would add it to vencord tbh
NICE
> saw electron patch note about colors
> immediately upgrades it on main
That was quick!
For me the problem started today and just a few hours after I notice it it is already fixed XD
This fixes this rare issue that occurs when having multiple screens on Wayland:
https://github.com/user-attachments/assets/a8908888-d9a6-440b-a79b-3a5a8aa82bee
bringing it over here, would it be possible for us to downgrade electron? It was upgraded last week in https://github.com/Vencord/Vesktop/commit/4e13f6d99632c52c35ab98e70e07dba1f014d080
An issue was raised here: https://issues.chromium.org/issues/476855843
Probably won't break other deps. Making it 39.2.7 fixes the issue, as it uses an older version of Chromium (142.0.7444.265)
vesktop settings missing from settings context menu ☹️
update your vencor
@vernal lintel yo not sure if this is relevant in any way but
I was checking journalctl logs to diagonose a crash of another app
and found this in yellow color:
fev 05 13:41:31 archlinux plasmashell[1135]: Trying to replace notification with id 1216 which doesn't exist, creating a new one. This is a bug in the application "vesktop"
hmm, seems like there's multiple of that:
https://paste.ononoki.org/?638b53900330c908#DjouMUqbShoLXF8uL1ZjSjEZQ4HuM5B1o7sT2V1xfgHC
DEPENDABOT YOU FUCKING PIECE OF SHIT I DO NOT GIVE SHIT ABOUT NODE TAR VULNERABILITIES
WHO CARESSSSSSS
it keeps spamming me with these irrelevant vulnerabilities
git gud
you suck!!
that reminds me of shellcheck warnings
titlebar is cursed again
X is not at the top-rightmost corner, it has an unneeded spacing there
yeah the .bar_c38106 has padding-right 12px but can't even determine where it comes from because devtools break
i didnt even notice
how do u guys notice this stuff
.bar_c38106 {
padding-right: 0;
}
this vesktop patch broke
because I tried to click the X from muscular memory and it didn't close the app
xd
it's always at the top-rightmost corner
funny part is that this also happens on the official discord client
vesktop is actually the app that does it correctly by patching it
haha
Somewhere between Discord app code and Chromium/Electron isnt checking if the notification is dismissed before trying to update it. not Vesktop
horrible, electron 40.x has a bug https://github.com/electron/electron/issues/49566
are you guys the mantainers of the vesktop packages on the AUR?
if so, I believe vesktop-bin (and maybe vesktop too) should be installed on /opt instead of /usr
as a reference, I have localsend-bin and github-desktop-bin from the AUR and they are both installed to /opt
https://wiki.archlinux.org/title/Arch_package_guidelines#Directories
Package files should follow these general directory guidelines:
/opt/pkg Large self-contained packages
hmm...
ah I see
since vesktop-bin depends on electron40-bin, I think it makes sense not to be on /opt
whereas localsend-bin and github-desktop-bin probably bundle Electron
also, maybe only recommend vesktop-bin in the website for a better user experience
nope
we do have comaintainer on vesktop-git
@humble mortar
why are we still recommending vesktop even
@humble mortar fix
Content
please. or some option in settings to make it enabled by default
time to update
any MacOs users in chat where the app never quits after using the updater
the main window closes but the app stays open
and i need to manually quit for the app to restart
@humble mortar fix
how can i even test updater without always downgrading to 1.6.4
Content
Since latest Discord update, I haven't been able to re-inject vencord. I keep getting the same error.
Content
Firstly, I'm sorry to open an issue here but I asked several times on the Discord and never got an answer.
I currently have an issue with idling in Vesktop (Linux): when in a audio channel, Vesktop will prevent swayidle from triggering completely.
This means that if I want to use my screensaver when going AFK, I have to disconnect from the channel.
You can test this easily by setting something like this in your sway config:
exec swayidle \
timeout 5 'ddcutil -d 1 s...
pretty sure this would have to be a vencord patch to block discord from getting wakelock
dont think theres a flag that disables it
or u can drop the request on the systemd side maybe
how does it get wakelock
theres like a zillion impls but they all use the dbus so just dbus-monitor while it does the thing
im gonna assume chromium uses the xdg portal
This PR adds the change from the referenced PR to Vesktop's arRPC patch, as the issue prevents RPC in Vesktop from detecting some games and the change has not yet been merged into arRPC
I have a question about the Arch package: vesktop-bin has electron40-bin (both packages from the AUR) as dependency. Isn't it a problem to update Electron separately from the main Vesktop client?
For example, official RPM packages bundle the client and Electron updates, whereas on Arch Electron can be updated independently.
@humble mortar
vesktop-bin isnt maintained by us and yes it can be problematic
@humble mortar COOKIE
dont you have perms or something
When venmic is active, if there's no input detected, does it automatically shut down the pipewire input? I noticed that it often cuts out the beginning of my sentences and i'm wondering if it's because it's taking time to create a new input when i'm very quiet. I am running easyeffects noise suppression so when i'm quiet, there's no audio at all. I don't have this issue with discord though.
is it bad to assume this is cause of the discord noise gate
venmic doesn't handle microphone input, only screenshare audio.
My assumption is this has something to do with the Discord client noise gate
all discord effects are off, and this issue only happens in vesktop, not discord or vencord on discord. Only other thing I can think of is the way the browser version of discord interacts with the microphone.
Content
Windows 11 25h2
Vesktop Last build
When enable the window transparency option it's break the title bar, I lost the buton to elarge or close the window.
I also notice that I can't drag the window.
Content
Title.
I believe that is the behavior on the official client, but I am not sure. Taskbar flashing is good due to muscle memory from Windows app flashing the icon when there's new shiny things to see.
And make it so it retroactively enables for everyone (new setting name) by default.
shows pipewire node descriptions instead of internal names (e.g. "Game Audio Output" instead of "output.loopback-2551-32") in the audio source picker
old:
new:
settings:
Really small change that makes it so if the Vesktop window is maximized it doesn't update windowBounds with it's new bounds, thus preserving the windowed state if the app is closed (restarted) while in fullscreen, making it work pretty much exactly like the official discord client.
Also noticed the minimized state wasn't being updated when "unminimized", fixed that aswell.
not even thanked :///
if you mean my PR, i did receive my thanks 😅
i disabled webhook for pr comments
because people post way too many replies with little relevance
i noticed you removed the license update, you have that on an eslint rule, is that only meant for new files ?
also ran into issues related to that due to most files being saved with CRLF
for every channel? or just #🖥-vesktop-development
yes
also ran into issues related to that due to most files being saved with CRLF
no files use crlf
if you have crlf, it's on your end and you need to configure your git right
hmmm, let me check again on my end, the repo is missing .gitattributes enforcing it though
not needed because formatter does it already https://github.com/Vencord/Vesktop/blob/main/.prettierrc.yaml#L8
.prettierrc.yaml: Line 8
endOfLine: lf
yup, that did it, that's very weird though because i always install it and configure it to checkout as-is 😅
anyways, there isn't a problem after all, thank you
watch out that's going to force LF in all git projects
it might not be what you want if you work with a lot of windows projects
but imo every git project should use exclusively LF so it's a good thing to set
yes, that is how i like it configured aswell, so it's all good 😄
you still have 2 files that are CRLF, but they aren't very important at all, but they managed to sneak in because there isn't a .gitattributes
.github\workflows\meta.yml
.vscode\settings.json
evil lewi
i made a little program to scan a given directory recursively for things i don't like seeing in my projects , for example the EOL being CRLF, EOF isn't a new line, and such, that's how OCD i am about it 🤣

sure man
just show me 30 irrelevant "vulnerabilities" in code that's not even used
why is arrpc being pulled from a specific commit, instead of npmjs registry ? I'm guessing because at some point in time there was a change to their code that was needed for Vesktop, but they have since released a new version 3.6.0
electron-updater is also outdated
ok
😅 i could do a PR for just that, but seems a bit pointless
ok, from what I'm seeing the electron-updater patch isn't needed anymore, they already wrap the check for the package manager in a try catch on the latest version, and they actually changed it from which to command -v too
https://github.com/electron-userland/electron-builder/blob/master/packages/electron-updater/src/LinuxUpdater.ts#L56-L63
and the one for arrpc is still valid, i can PR if you want, but not that much changed tbh
LinuxUpdater.ts: Lines 56-63
protected hasCommand(cmd: string): boolean {
try {
this.spawnSyncLog(`command`, ["-v", cmd])
return true
} catch {
return false
}
}
there is 0 reason to update
i'd say getting rid of a patched dependency is a pretty good reason, but if you don't want to, then it is what it is 🤷♂️
Fixes vencord taking over the nested desktop session, as described in [this discord message](#🖥-vesktop-support-🖥 message).
when running npm test, the linter requested the statement be split into multiple lines, so I have done so.
While testing, I noticed that the XDG_CURRENT_DESKTOP variable is actually also set by gamescope, so I changed the discussed check to pass on XDG_CURRENT_DESKTOP === "gamescope". Any desktop env...
fixes handleSync overriding e.returnValue in some cases
also updates the spellcheck languages handler to return the value
instead of mutating the event, which could lead to undefined on the renderer side
aa6ebc4 Bump electron from 41.0.4 to 41.1.0 - dependabot[bot]
Bumps electron from 41.0.4 to 41.1.0.
Release notes
Sourced from electron's releases.
electron v41.1.0
Release Notes for v41.1.0
Features
Added nativeTheme.shouldDifferentiateWithoutColor on macOS. #50408 (Also in 42)
Notes: Added support for the urgency option in Notifications on Windows. #50382 (Also in 42)
Fixes
Fixed a bug where Windows notification icons could fail to save because their temporary filenames contained invalid characters. #50483 ...
shush
d179b62 implement global shortcuts with electron api - Vendicated
works on kde but doesnt work on gnome

it should work on gnome but electron doesn't even send the portal call 
i blame electron/chromium
it just returns success=false and no error, very useful
insane
how does electron even handle it does it check if the portal supports or is it hardcoded to work on kde and not gnome or smth
idk
I think it checks if wayland or x11
yeah I can't find anything useful on how it determines that and I don't exactly feel like grepping the entire chromium source for that but if it works on kde not on gnome I'm gonna guess it just has a list of working portals and doesn't even try if one of them isn't detected
I'll just add a way to invoke shortcuts manually
eg via vesktop --shortcut toggleMute
then people can add desktop shortcuts that run that command
sure
every desktop has a way to register custom keybinds that run commands
wasn’t there a pr for global shortcuts @vernal lintel
this works but it takes like 2 seconds
sucks ass
overcomplicated and unmaintainable tbh
how
it wont solve the issue
i will probably use a unix socket server
echo "toggleMute" | nc -U $XDG_RUNTIME_DIR/vesktop.socket
smth like that
true
how do u plan to handle x11/bindows
or does the electron api do that
I thought about it and it's not feasible for me to add such a large library + maintain it long term in a language I am not proficient at
it does
my own implementation using electron api is much simpler and more maintainable
doesnt support push to talk and mouse buttons but so be it
thats why I'm adding an alternative way to trigger shortcuts so people can do their own shortcuts via desktop
KDE/Hyprland users test this branch please thanks
9d7ca67 Add ipc method to invoke keybinds - Vendicated
This adds global shortcuts for the following actions: toggle mute, toggle deaf, toggle streamer mode, disconnect from vc
Limitations:
- No push to talk, use toggle mute instead
- No mouse keys
- On Linux Wayland, only works on some compositors (KDE confirmed working, GNOME confirmed not working)
As a workaround for the third limitation, you can also trigger shortcuts via a unix socket or vesktop cli flag. However, the cli flag has a delay of ~2 seconds so the socket is rec...
This optimizes SOME of the files here, overall with lossless optimization.
The exception is to build/icon.svg and the webp
The webp has had timing of it slightly adjusted, though the frames themself have no lossy changes otherwise.
The png use https://github.com/oxipng/oxipng via https://ezgif.com/optipng and the other content is optimized through similar means
The svg though is instead optimized via https://svgomg.net
some new electron builder version somehow messed up the dmg
Content
Bug Description
Window transparency stops working after rebooting the system. It only works during the session where it was first configured, never again after a reboot.
Steps to Reproduce
- Install Vesktop 1.6.5 (.deb) on KDE Neon
- Enable Transparency in Vesktop Settings
- Confirm transparency works
- Reboot the system
- Open Vesktop again
Expected Behavior
Transparency should persist across reboots since transparent: true is correctly saved in settings.json
...
vesktop fixed finally
vibesktop fixed, it is now fully human
Summary
- keep the selected Linux screenshare profile aligned across capture constraints,
videoStreamParameters, and RTP sender parameters - resync the active screenshare path after stream start, profile changes, and
replaceTrack() - add a guarded recovery path for viewer attach/rejoin when the outbound sender still appears stalled after the normal resync path
Context
This started from Linux NVIDIA VAAPI/NVENC validation, especially Prefer Clarity sessions.
Before this patch, th...
his clawdbot found vesktop and wanted to contribute
no human is making a 1k long file and pring it without at least discussing here or mentioning some issue
holy yap


✍️ Bug Description
No Audio Sources show up when I go to Stream in Vesktop.
I open up a video in VLC Player, go to stream it and there is no option to select an Audio Source.
It doesn't matter if I pick Screen or Window, nothing changes.
I've tried (I think) every combination of audio settings and nothing seems to appear.
I've tried reinstalling everything from scratch and it hasn't helped. (Including themes/plugins etc.)
Let me know if you need any more info, thanks!
🔁 Ste...
Bumping this, having the same issue right as I'm trying to stream a game with my partner.
According to the logs, venmic properly detects various nodes, which also includes VLC:
[2026-04-15 21:35:57.578] [venmic] [trace] [patchbay] (handle) new node: 90 (name: "VLC media player (LibVLC 3.0.22)", app: "VLC media player (LibVLC 3.0.22)")
Maybe this is a vesktop issue?
@vernal lintel i'm not at my pc rn, i can check the next few days (will also update venmic to latest rohrkabel), but according to the logs this looks more like a vesktop issue
not venmic issue
Nice
Btw, Ive recently experimented with zig (as a compiler) again for saucer and actually got it working (had some issues before due to outdated libc++ and llvm backend) but it now works quite well and I could properly target older libc
Not sure if thats still a concern but I'll probably also include the necessary setup in the repo so that we can technically target older libc for those that might need it
Confirmed to not be a venmic issue
I've checked the Vesktop Discord and it seems like a widespread issue:
<img width="686" height="564" alt="Image" src="https://github.com/user-attachments/assets/080ea8da-87dc-49f3-865b-b02171335fdc" />
Thanks for confirming it looks good on this end, sorry for the trouble
I've checked the Vesktop Discord and it seems like a widespread issue:
<img width="686" height="564" alt="Image" src="https://github.com/user-attachments/assets/080ea8da-87dc-49f3-865b-b02171335fdc" />
Thanks for confirming it looks good on this end, sorry for the trouble
No problem, thanks for reporting!
why not js leave it open
aaaa
Content
I can not see Krisp In Noise Suppression.
I use Ubuntu 24.04.4 LTS , Linux.
I updated the newest version of discord
This commit improves code quality and maintainability through several refactoring changes.
Key improvements:
- Extract shared
buildLinkDatahelper function from duplicateVIRT_MIC_STARTandVIRT_MIC_START_SYSTEMhandlers - Replace unsafe
catch (e: any)blocks with proper TypeScript error handling using type narrowing - Remove untyped
as anycast on GitHub Authorization header ingithubGet() - Tighten
once()generic type signature for improved type inference on Paramet...
well atleast they were honest
AI’s append their description with that I think
Content
plis update electron persion for this app, Versions prior to 40 experience what I describe in the title.
Is there any specific reason why that wouldn't be satisfactory?
While toggle mute does meet my use case, I simply prefer using push-to-talk. I like to appear unmuted if I'm available to talk. The sound when you mute and unmute is longer and different. I think it would be best to offer as close to a vanilla experience as possible, so I think that some people would prefer the option to push-to-talk (with some kind of external setup) rather than just toggle mute.
Any progress on this? Im on NixOS and I have the same issue. Screen share becomes unwatchable.
Also, push to mute is usful when dealing with a in-game voicechat and being in a unrelated discord voicechat
@Svnipni Did you ever find any solution to this? Last I checked I heard that Chromium fixed this but it seems to not be working or something?
https://issues.chromium.org/issues/399412221#comment19 mentions 144.0.7543.0, but my Vesktop is running on 144.0.7559.134 and it's still going on.
Because most users don't need to use non-stable branch
ai
pai
Content
with 1.6.4
PPIMAGE env is not defined, current application is not an AppImage
Vesktop v1.6.4
Disabled Chromium features: WinRetrieveSuggestionsOnlyOnDemand, HardwareMediaKeyHandling, MediaSessionService
Gtk-Message: 21:01:38.248: Failed to load module "appmenu-gtk-module"
[13933:0424/210138.321026:ERROR:dbus/object_proxy.cc:573] Failed to call method: org.freedesktop.systemd1.Manager.StartTransientUnit: object_path= /org/freedesktop/systemd1: org.freedesktop.systemd1.UnitExists:...
Outdated vesktop + vencord - use our support channel
Fix for https://github.com/Vencord/Vesktop/issues/464
This change ensures the window's state is restored before handling the FOCUS event, preventing the incorrect window sizing behavior.
Unfortunately, I was not able to test how this behaves, or whether it causes any issues on non-Windows systems.
Summary
When Vesktop's renderer process dies (commonly during long screen-share sessions, see #792), the main Electron process keeps running but the window appears frozen because there's no UI to update. The existing render-process-gone handler only logs the event:
mainWin.webContents.on("render-process-gone", (event, details) => console.log(details));
This PR makes that handler reload the renderer when the failure looks recoverable, so the user gets a brief interruption ins...
Opened #1249 as a partial mitigation: when the renderer dies during a screen-share, auto-reload it instead of leaving the window frozen.
It does not fix the underlying cause (the renderer being overwhelmed on NVIDIA + Wayland during high-resolution capture is upstream Electron/driver territory), but it converts "stream silently turns into a dead window after 20–30 minutes" into "stream blips for a couple of seconds and recovers."
Skips auto-reload on clean-exit / killed / launch-failed / in...
I've been having this issue when sharing my entire system audio, other users in the call can hear themselves in my stream. I've been having this issue in both vesktop/equibop but also fluxer (which recently integrated venmic for sound sharing).
I'm not sure it's a bug so im not opening a bug issue but i don't know how to fix this. What's also weird is that it suddenly starting doing this, it wasn't doing it before
Venmic has an exclude options where clients should include themselves as to not share theirselves during entire system screenshare. Perhaps the clients are setting their exclusion improperly?
Venmic has an exclude option where clients should include themselves as to not share theirselves during entire system screenshare. Perhaps the clients are setting their exclusion improperly?
Are u talking abt this exclude source dropdown ?
<img width="593" height="436" alt="Image" src="https://github.com/user-attachments/assets/e2f81238-7926-4f8f-bce5-6808bac03c30" />
Venmic has an exclude option where clients should include themselves as to not share theirselves during entire system screenshare. Perhaps the clients are setting their exclusion improperly?
Are u talking abt this exclude source dropdown ?
<img width="593" height="436" alt="Image" src="https://github.com/user-attachments/assets/e2f81238-7926-4f8f-bce5-6808bac03c30" />
No - I'm refering to the API, clients are supposed to find their own node and then add it to the exclude list d...
No - I'm refering to the API, clients are supposed to find their own node and then add it to the exclude list directly.
Well that's odd, that would mean none of the clients are doing correctly, aka vesktop, equibop (which is a fork of vesktop) and fluxer. I'm genuinelly so confused of why that is, is there any way i could fix that ?
ok so apparently i had easyeffects running in the background and it fucked up everything lmfaoooo well ig it's fixed now thanks!!!
This PR introduces token protection to mitigate common token grabbers that extract credentials directly from disk (e.g. LevelDB or Cookies SQLite).
Overview
When Electron's safeStorage is available, Discord tokens are no longer stored in plaintext. Instead:
-
localStorage (
token)- Intercepted via a proxy
- Stored as an encrypted blob under a separate key
- Decrypted transparently at runtime
-
Cookies (
tokens)- Intercepted via a
document.cookieproxy
...
- Intercepted via a
I've done this, because I am a server mod and one of our mods got hacked by this. So I wanted to prevent that easy way of token grabbing. My test code is the following. Executed inside ~/.config/vesktop/sessionData/Local Storage/leveldb
import plyvel
db = plyvel.DB(".")
for key, value in db.iterator():
if "_https://discordapp.com\x00\x01token" == key.decode("UTF-8"):
print(value.decode("utf-8"))
```
Why does the copyright note say 2023
Is anyone else having a problem with spell check not working after deps bump commit?
Pre it works no prob
dont send ai generated pull requests.
Ah yes, typical incompetence. Anyways, have a great day and continue with getting discord token stolen
Incompetence is not wanting aislop garbage prs? Sure, man
yea i noticed this too, not on vencord tho, but equibop
it works fine for me
@zspher I think I found a solution (at least for Vesktop) that might be interesting to both https://github.com/Legcord/Legcord/issues/1011 and https://github.com/tailscale/tailscale/issues/10396 too.
I think it works anyway.
Basically if you add a call to win.webContents.setWebRTCIPHandlingPolicy("default_public_and_private_interfaces");, it seems to resolve the issue.
It no longer checks the tailscale0 interface, but should still support internet VPNs like Mullvad, Surfshark or whatever el...
I think the issue is that when you have tailscale, ICE tries to make a connection over it, but the response comes back on the normal network, which confuses ICE and gets it into a broken state. Just stopping it from using non-default routes seems to fix it.
This changes the WebRTC IP handling policy to ignore non-default routes.
When running Tailscale, WebRTC can sometimes get confused and bind to that private VPN interface, hanging forever on DTLS Connecting.
This change, from my understanding, makes WebRTC only bind to the default route interface which is likely the intended behaviour anyway, avoiding weird issues with tailscale.
Can the default discord's built-in priority push-to-talk also be considered natively?
I think the issue is that when you have tailscale, ICE tries to make a connection over it, but the response comes back on the normal network, which confuses ICE and gets it into a broken state. Just stopping it from using non-default routes seems to fix it.
how could i stop it from using non-default routes?
i've been having this issue of "DTLS Connecting" since i started using vesktop and could never figure it out
my pc is connected to a zerotier network with my partner
Running CachyOS and KDE 6.6.4 I did try it on my system and I noticed an odd behavior when I set both toggleMute and toggleDeafen:
After registering both shortcuts toggle mute only allows me to undeafen (does not toggle). But after I trigger that action by clicking the button at least once it starts working (but it still undeafens as well). That problem reappears every time I restart vesktop. Seems to happen no matter if I use KDE shortcuts or run a script through the terminal.
Content
Hello - A recent update seems to have broken the tray notifications indicator.
Summary
On non-GNOME Wayland compositors (niri, hyprland, sway, labwc) Vesktop currently shows the xdg-desktop-portal chooser dialog twice in a row when a user clicks "Share Screen":
- First chooser: triggered by
desktopCapturer.getSources()enumerating sources for Vesktop's preview thumbnails - Second chooser: triggered by Chromium's
WebRTCPipeWireCapturer.Start()whensetDisplayMediaRequestHandler's callback fires — this opens a fresh portal session independent from the fir...
screen sharing not working on Plasma.
For some wierd reason I have this bug when streaming any browser but firefox in private mode
when did vee unmute comments for vesktop repo
On windows when enabling the window transparency the window has to be frameless (without the title bar) to work.
So on windows there is currently no way to implement the transparency without losing the title bar.
Are you using windows? If so then the taskbar icon is flashing but 4 times. This is hard coded inside the chromium kernel so there is no alternative for an infinite flash if you want that.
No, I am on Linux. IIRC, Windows apps can flash up to seven times in the taskbar.
@MatthewCash not sure if you're still interested in this, but I've started working on this again and I've pulled your commit into my fork, hopefully that's okay with you. I haven't fully tested/recompiled your implementation yet as I've been busy reconstructing the powerMonitor events and basically remaking the plugin from scratch but I thought I should probably keep you updated on this.
That can be true on native apps but Vesktop is based on electron which uses a chromium browser. So the OS specific attributes are not considered here.
Currently I don't have the ability to test on a linux system so I can currently not further investigate this.
Does the flashing setting not work at linux which is under "Vesktop Settings"?
<img width="765" height="91" alt="Image" src="https://github.com/user-attachments/assets/eb89bee1-a66c-44d5-944d-9baf30a87b83" />
It does work. My request is to enable that option by default. It's OFF by default, regardless of platform:
https://github.com/Vencord/Vesktop/pull/1254
If the PR is accepted then the flashing is on by default.
it's off by default..?
Sorry but I have no interest in doing this. Windows users should just use the normal Discord app for streaming
Add button in browser popup to be able to also open the site inside the browser
https://github.com/Vencord/Vesktop/pull/1256
This should be like the feature you wanted :D If there are still other opinions about the implementation tell me ^^
The commit is fully working, there is no work to do.
I don't really have time to work on this anymore so I'm just hoping you can get this fixed :)
@LucaHDR That fixes the issue, but only partially. To make it default true, a new setting name has to be defined, so it retroactively enables the option for everyone. Otherwise, only new installs would be affected.
I understand that but imo it would also be annoying to users that did disable it and then it magically enables again.
Well, it was disabled by default, it would only hurt people that: enabled it; didn't like it; disabled it again.
However, we can't differentiate those people from the people that didn't ever enable it in the first place. There's no code that stores how many times you toggled the setting or anything like that, as far as I am aware.
So it's a bit of a necessary evil in this case. It would be communicated in the changelog, of course, so people would know about it.
I am worried if this works o...
Okay I can also change that. Could take some time cuz when I am back home I have to get an overview how the settings exactly work (I am since 2 days actively working for Vesktop). So the apeed of changes could be slower than expected.
I made a pull request #1255 that fixes it
@next sundial Sorry, but I'm blocking you from leaving further comments. You're regularely leaving excessive, overly opinionated, demanding comments where you act like you're a maintainer and your opinions carry weight
It's annoying
I did not ask LucaHDR to make a PR. He decided to join the issue and discuss about it with me. If he wants to make a PR and make changes voluntarily, that is his choice. The issue has been up since 17 March; you or any other maintainer could have closed it beforehand. It is also possible to see that I do not have the "Member" tag when I comment, meaning that I am not a maintainer.
The fact that I am using correct English grammar to express my ideas/intentions clearly does not mean I am demanding something or that I do not think other opinions are valid. In fact, I am completely fine with the request being rejected.
I believe it is an over-reaction from your side, considering I do not frequently make issues/comments in the projects' repositories. But that is your choice and I respect it.
well, the first one is because I was mentioned in the PR, and I really don't like to read things that seem to be vibecoded. the description of that PR is very non-natural and seems like an AI wrote it
so I just commented that I don't want to talk in order not to appear that I am ignoring it, and that's it
the second one is just what I said about using correct English grammar
but, again, I am fine with your decision, even though I believe it is an over-reaction
drink water and be happy 👍
hum, I just saw that the first PR was closed without discussion
welp, poor contributor
https://github.com/Vencord/Vesktop/issues/1252
I cannot reproduce this issue locally
Content Hello - A recent update seems to have broken the tray notifications indicator. This is all that's seen regardless of read or unread messages. There used to be a dot.
works both on taskbar and tray for me
my setup is on Arch Linux, with vesktop from the AUR on KDE Plasma 6.6.4
I’m testing electron updater with different identities rn
Same by default in gnome, no more close, minimise button
I’ll make a pr soon with these changes
I just need to quadruple check everything so nothing goes wrong
If it does go wrong then Mac users will need to manually update
As of May 6, 2026, this bug seems to be back.
didnt even wait for a response #🖥-vesktop-support-🖥 message
Repro
- Open Vesktop on macOS.
- Use Vesktop > Force Update Vencord.
- Hit a Vencord asset download failure.
- Relaunch Vesktop.
On my machine, this left ~/Library/Application Support/vesktop/sessionData/vencordFiles with only package.json. Startup logs then retried vencordDesktop*.map and .LEGAL.txt downloads, and Vesktop stayed alive with no main window.
Change
downloadVencordFiles() now maps release assets by exact filename and downloads only the four files `isVal...
Why did you close the issue, are the reproduction steps not clear enough? I actually had to fix this myself because my local vesktop install broke, and not even reinstalling the app using homebrew fixed it. Sure, I’ll give you that the pull request looks like spammy made up LLM issue, but I assure you I verified everything myself
because it's aislop and stupid
You're solving a made up issue that does not actually exist with your shitty ai code
Literally peak ai brainrot, your brain is smoosh from all the AI you fed it
oh my god
Content
When I share my screen with someone, they hear all the sounds including themselves.
This happened to me today, spent a while trying to fix it before I launched from terminal and that fixed it, so the icon shortcut works again.
Any idea why that happens?
girl im bored lets delete the vesktop repo
Today I learned that I am not built for c++. I am a spoiled child and I need my cargo and my compiler and my docs.rs. 4 hours of headaches and I still can't get a destructor working for IdleNotifier. I'm probably just going to put back my rust implementation with proper integration into the buildsystem (with docker and build scripts for prebuilts and the likes), unless somebody else who is actually capable and has free time can patch things up.
I'll wait till the end of the week before doi...
close random issues and flame argumenters
Related: https://github.com/Squirrel/Squirrel.Mac/issues/160
This will whitelist a new signing certificate that will be used for signing Vesktop, so there's no issue with electron builders auto-updater (Squirrel.Mac).
Since https://github.com/Vencord/Vesktop/releases/tag/v1.6.5 only allows updating to versions signed with the same identity, we will need to create a new release with that same identity for it to update to (with the whitelist), t...
I made an image bc im bad at explaining
@sly cairn the cert change can’t happen immediately because it breaks the updater (users would need to manually update to the new version) but the release after the next release it can be changed

okie dokie
im actually getting a mac soon so i can help with mac development in some capacity
though cuz its a business machine i can only do a small amount

huge honestly
I don't really know much typescript to help that much with vesktop for mac
dw Claude can do it
pixelmator pro
Can confirm identical issue present on Windows 11 as well
Found this issue looking into why the status was inconsistent compared to the official client. Im trying a plugin for now, to see if it can help. Im on X11.
<img width="884" height="291" alt="Image" src="https://github.com/user-attachments/assets/c7798bc7-19fa-4ed4-a16b-e07f6530a9be" />
idk if I sent this here already but this is vesktops mac icon project
This appears to be fixed for me now, not quite sure how soon but it seems it's repaired itself. Might have been related to me clicking "repair vencord" when right clicking the tray icon.
Did discord remove that option from the streaming window? I wanted to look into that but in the normal client I can't find the checkbox for the notification when starting a stream
I suppose you have seen https://github.com/MuffinTastic/wayafknext-monitor?
It appears to handle both Wayland and Mutter sides of things.
I can try to find some time to help with Rust implementation.
I'm going to be completely honest, I've never seen that before. I was talking about the prior package I made (you can see it here) which Matthew replaced with a direct libvesktop addition due to concerns about reproducibility and lack of proper integration. I'm pretty sure I can set up something like another dockerfile build script to pin rustc/cargo properly for reproducibility on p...
Is there any specific reason why that wouldn't be satisfactory?
Channels which require push-to-talk.
Vencord is still broken, but discord canary looks like it has been working for me.
If you want to force this to happen, though it's a bit involved.
On Windows OBS allows you to isolate individual program audio, as well as create a virtual camera.
On linux you need to install the linux-pipewire-audio plugin (to isolate audio) as well as the v4l2loopback kernel module and enable it with the exclusive_caps=1 option (to create the virtual camera).
Once you've done that, you can just select the camera option in discord, select the virtual camera, and stream with whateve...
Are you using the discord desktop client or the discord web client? Afaik Vesktop is a wrapper for the web client, so the desktop version, a native electron app, is not comparable to Vesktop.
On Fri, May 8, 2026, at 04:42, Eliott wrote:
Yot360 left a comment (Vencord/Vesktop#528) https://github.com/Vencord/Vesktop/issues/528#issuecomment-4405396342
Vesktop is still broken, but discord canary looks like it has been working for me.—
Reply to this email directly, view it ...
@SlyCedix I think that the main issue currently is about maintainability. A solution has been proposed via a PR from @quadrinex but the code for it is a bit much. The code is primarily in C++ while most of Vesktop is written in TS / JS. It also doesn't help that @Vendicated looks to be primarily on Linux, making it harder or more annoying of an issue to test and maintain if / when bugs relating to the C++ code from the PR gets submitted.
While sure, the virtual camera workaround looks to...
Oh certainly, having everything work out of the box is ideal. I just had already found a bandaid for the problem so I wanted to share it. I was actually here to look at a different issue.
May 8, 2026 11:59:58 arion @.***>:
[Image]*JustArion* left a comment (Vencord/Vesktop#569)[https://github.com/Vencord/Vesktop/issues/569#issuecomment-4407845536]@SlyCedix[https://github.com/SlyCedix] I think that the main issue currently is about maintainability. A solu...
I didn't find any way to make it lighter because electron doesn't support capturing an app audio natively, the solution I used is actually (I think) the same that obs uses to capture an app audio. I may search on making it lighter but i'm pretty sure its not possible to make it in TS / JS for now.
@SlyCedix A solution has been proposed via a PR from @quadrinex but the code for it is a bit much. The code is primarily in C++ ...
I wonder if using something like deno's FFI would allow implementing this feature in JS/TS. If that's built as a standalone app, then there is not much issue with running separate runtimes.
https://docs.deno.com/runtime/fundamentals/ffi/
May 8, 2026 17:38:30 Quadrinex @.***>:
[Image]*quadrinex* left a comment (Vencord/Vesktop#569)[https://github.com/Vencord/Vesktop/issues/569#issuecomment-4410059060]I didn't find any way to make it lighter because ele...
The issue isn't language or whatever. It's the fact that this adds additional complexity and something I will have to maintain (I don't even use Windows), for very little gain.
Screenshare on the official Discord app is great on Windows, Vesktop won't ever come close to it. Considering this fact, it'd be a wasted effort just to have an inferior alternative to an already existing solution. I consider this out of scope.
We provide the most basic Windows implementation for people who are okay ...
Not really. FFI is fine for tiny WinAPI calls, but not for the actual WASAPI process-loopback capture.
That part needs async COM objects, callbacks, native audio buffers, event handles, and careful lifetime/thread handling. Doing it through JS FFI would be fragile and crash-prone, and you’d probably end up writing a native wrapper anyway.
I wonder if using something like deno's FFI would allow implementing this feature in JS/TS. If that's built as a standalone app, then there is not much ...
Would a way to pick a camera as a source for the stream from a camera be a good compromise?
If someone from the community wants to make a nice easy companion app for it, that's great. If not, people can use the OBS workaround if they want.
Eventually, maybe a way to launch the application from vesktop later down the line, once it's matured.
May 8, 2026 18:38:09 V @.***>:
[Image]*Vendicated* left a comment (Vencord/Vesktop#569)[https://github.com/Vencord/Vesktop/is...
just close as not planned tbh
For OBS you can right click on the canvas and click "Open Preview Projector" > "New Window"
Tho a good thing would be to change the audio source to a virtual mic
Would a way to pick a camera as a source for the stream from a camera be a good compromise?
If someone from the community wants to make a nice easy companion app for it, that's great. If not, people can use the OBS workaround if they want.
Eventually, maybe a way to launch the application from vesktop later down the line...
I like the idea of using obs camera
They suddenly work with normal discord now for me
I gave up on Vesktop ever having this properly fixed, this issue should have been reopened 5 months ago.
Since I've never had issues with the official Flatpak, I now make my own Flatpak from the official manifest, and I add a build step that runs the Vencord CLI before packaging. Badges work and everything. If the official Flatpak is broken for you then there must be something wrong with your KDE configuration somehow 🤔
I have the official Flatpak and the badges work without Vencord, I'm not using Vesktop.
Pull request overview
This PR updates the macOS build/signing configuration to “pivot” the app’s designated code signing requirement so Squirrel.Mac-based auto-updates can transition from the old signing identity to a new one by allowing both during an interim release.
Changes:
- Configure
electron-buildermac signing to use a custom requirements file. - Add a designated requirement expression intended to allow both the old leaf certificate (by SHA-1 hash) and the new Team ID / Dev...
The designated requirement here does not constrain the app identifier (bundle ID). When Squirrel.Mac compares designated requirements for updates, a requirement that only checks the signing cert/team can become overly broad and may allow an update payload signed by the allowed identity but with a different identifier. Consider including identifier "dev.vencord.vesktop" (matching build.appId) as part of the designated requirement, and keep the cert allowlist logic inside that identifier/an...
- Electron 41.5
- Tightened engine fields
- pnpm 10.33.3
- electron-updater 6.8 -> drop patch
you can't add yourself to devs on vesktop
@compact fjord you have lost it...
I'm not seeing my camera appear as an option. (I'm on Linux)
OS: Manjaro Linux x86_64
Kernel: Linux 7.0.3-1-MANJARO
Running your fork via pnpm start
Okay I added push to talk in two ways: toggle and start/stop. You can either use it as normal keybind in app and use toggle/two buttons, or you can run it from an external process if you want proper push to talk
<img width="748" height="680" alt="image" src="https://github.com/user-attachments/assets/685e1a2e-d2d2-4de2-9f11-9e3b58dcb2a7" />
I also added decoupled mute/unmute options
Let me know if these work well for you. Also @Malix-Labs please see if the priority PTT works for you
Very cool, imma test that this week
I tried testing it using
nix run --impure --expr '
let
pkgs = import <nixpkgs> {};
src = pkgs.fetchFromGitHub {
owner = "Vencord";
repo = "Vesktop";
rev = "797c02bdd0d35063ac76d75cffa77b0790f3a657";
hash = "sha256-xeeg1U0+1lfJxTZQsY5Y49aM66Qp/1zqVkKSiewSxJk=";
};
in
(pkgs.vesktop.override { electron_40 = pkgs.electron_41; }).overrideAttrs (_: {
inherit src;
pnpmDeps = pkgs.fetchPnpmDeps {
pname = "vesktop";
...
Update:
It does trigger the global shortcut portal but only after I closed and relaunched the app, and only for a single shortcut
<img width="1795" height="1052" alt="image" src="https://github.com/user-attachments/assets/bcd05de0-df39-4e36-884c-fd7fc5bd4004" />
And after accepting the prompt, it globally works
However, when I did go into KDE Plasma's settings and deleted the global shortcut, they still were showcased as if they were registered in Vesktop.
Also, it's not possib...
Also, looks like it's still "electron" that is asking for keybindings, and not "vesktop"
That's probably just a nix run issue (.desktop doesn't integrate well with nix run)
Can you check on your side?
Rust binaries are usually reproducible OOTB, although there are some edge cases. I don't have experience with napi, but chances are it might be reproducible already. Making it easier to build for non-Rust people might be sensible though.
Given that Gnome's Mutter is probably the only popular compositor missing ext_idle_notifier_v1, it sounds like we should probably handle it when already considering zbus for KDE's lockscreen.
I won't be able to look into it before the next week.
I mi...
I'm wondering how relevant all of this is now, considering Discord's latest announcement video regarding Linux:
https://youtu.be/BwNfmazmU4o?t=184
We also fixed a bunch of smaller issues. [...] Wayland idle protocol support, which means you can step away from your desktop, and still get notifications to your phone. [...]
And indeed, I've heard from some friends who don't use Vencord/Vesktop (and by extension don't use my plugin) that they've noticed their idle detection has started...
Well that's great news for people who use the official client and are willing to switch to it, but I personally don't use the official client and probably won't as long as Vesktop's an option for me, regardless of how good its linux support becomes in the future. Idle detection only works on the official client itself regardless, so Vesktop is completely unaffected by this anyways.
I just wanted feature parity with the official client, and if this is a recent update that means this PR was ...
Summary
discord:// invite links are dropped on the floor when Vesktop is already running — the window focuses but no join modal appears. This fixes that for both Linux/Windows (via second-instance) and macOS (via post-startup open-url).
Background
#813 added discord:// URL handling, but only at first launch:
createMainWindow()readsprocess.argv/darwinURLand callsloadUrl()once.- The
second-instancehandler insrc/main/index.tsignoredcmdLineentirely (...
vchars —
— EM DASH
u didnt even need to check that to see if it were ai btw
oh yk
im experienced with ai i would know yk
Idle detection only works on the official client itself regardless, so Vesktop is completely unaffected by this anyways.
Ah yeah, true. Disregard my previous comment...
4 hours of headaches and I still can't get a destructor working for IdleNotifier.
I wasn't familiar with Napi before reading through the drafts here, but Napi's ObjectWrap API immediately caught my eye. It seems as though the lifetime for objects created through that API are fully controlled by V8, so for the dest...
I’m completely out of my element with C and C++, so unless someone is willing to fix and extend libvesktop to encompass everything everyone has spoken about, I’d still prefer to put the rust module back in place since it’s something I can at least work on myself. Regarding destructors and the IdleNotifier objects, It’s pretty sure on wayland it’s always there once loaded by the plugin, but only one is loaded per Vesktop instance. Basically, I don’t see a point in a dedicated stop/start system...
Crazy
Just a note, (not sure if anybody cares but maybe @mati865 does?) I'm going to probably force-push/reset this branch
No worries.
Well that's great news for people who use the official client and are willing to switch to it
Yeah, the ship has sailed for me as well. Moved fully to Vesktop last week and removed official app 3 days ago.
I have been using a patched version of vesktop with this PR for a while now and it works perfectly
This fixes the main Discord window, but it misses Discord pop-out windows. Electron exposes setWebRTCIPHandlingPolicy on WebContents, and Vesktop creates Discord popouts as separate BrowserWindows, so those popouts get their own webContents.
An idea is that this be centralized in a small helper and applied both to mainWin.webContents and to Discord popouts in setupPopout? Otherwise voice/stream popouts will still hit the DTLS Connecting issue when a VPN interface like Tailscale...
have there been thoughts of using capacitor?
so that vesktop, since its electro, will run on ios and android
you are losing it
i am curious as to why it wont work tho
ik from a frnd that forked hayase that its possible
he doesnt have it working fully yet but it runs
i mean, it could work as web app, with ofc tweaks to the site that would make it look/work as the actual app
ah fuck it lemme js fork and try and see
best mindset to have honestly
(im going to fail, i barely know javascript)
Well like, if we wanted this then we would just use our browsers
There’s a reason discord desktop is for desktop
Review written by ChatGPT.
Review written with ChatGPT.
This sets the application menu globally, but I don't see any matching restoration of the default menu when the popup closes or when the popup URL is no longer relevant. After opening one popup, the app can keep an "Open in browser" item pointing at a stale URL until another menu update or restart. Please restore the default menu when the popup closes, or avoid modifying the global application menu for popup-specific state.
Review written with ChatGPT.
"initBrowserPopupMenuBar(win, url)" runs before the URL is parsed/validated and before this popup is confirmed to be the in-app browser popup case. Since "Menu.setApplicationMenu()" mutates the global application menu, denied/invalid/non-browser popup URLs can still replace the app menu. This should probably only run after the handler has confirmed that the popup is allowed and handled as the browser popup.
Review written with ChatGPT.
This seems like the wrong abstraction layer for a popup-specific action. "Menu.setApplicationMenu()" affects the entire app, not just the popup. A popup-local UI button, context menu, or window-specific menu would avoid leaking popup state into the global app menu.
wtf 😭
When sharing audio from an app whose audio plays through a child process (e.g. Rhythia's FMOD subprocess), only the directly selected node was linked, so the actual audio never made it into the stream. Walk PPid via /proc//status to group sibling audio nodes under their topmost audio-owning ancestor and expand include/exclude filters to cover the whole group. Skipped when granular selection is on.
how do you use ts im stupid and js want my deafen bind
f4866ca Bump axios from 1.7.9 to 1.16.1 - dependabot[bot]
Bumps axios from 1.7.9 to 1.16.1.
Release notes
Sourced from axios's releases.
v1.16.1 — May 13, 2026
This release ships a defence-in-depth fix for prototype pollution in formDataToJSON, hardens proxy and CI workflows, restores Webpack 4 compatibility for the fetch adapter, and includes several small bug fixes and maintenance improvements.
🔒 Security Fixes
Prototype Pollution Defence-in-Depth: Hardened formDataToJSON against already-polluted Object.prototy...
OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting @dependabot ignore this major version or @dependabot ignore this minor version.
If you change your mind, just re-open this PR and I'll resolve any conflicts on it.
1.7 to 1.16 😭
cant be a bump if its a downgade
Summary
- Fix screenshare timeout: Reduced thumbnail sizes in the screen picker flow to prevent Discord from timing out while waiting for the Electron media handler callback. Wayland picker thumbnails reduced from 1920x1080 to 320x180 (auto-selected source doesn't need a high-res preview). Large preview thumbnails reduced from 1920x1080 to 1280x720.
- Fix audio stream memory leak: Audio tracks from
getUserMediawere removed from the stream without calling.stop(), leaving nativ...
This has been tested by the submitter on their local Vesktop Flatpak setup — screenshare no longer times out, and no memory growth was observed after multiple start/stop cycles.
Hope this helps, have a good day! 🙂
Don't send garbage aislop.
the audio track part is probably valid but the rest is just absolute nonsense
88a0d23 Bump brace-expansion from 1.1.12 to 5.0.6 - dependabot[bot]
Bumps brace-expansion from 1.1.12 to 5.0.6.
Release notes
Sourced from brace-expansion's releases.
v4.0.1
fmt 5a5cc17
Fix potential ReDoS Vulnerability or Inefficient Regular Expression (#65) 0b6a978
https://github.com/juliangruber/brace-expansion/compare/v4.0.0...v4.0.1
v4.0.0
feat: use string replaces instead of splits (#64) 278132b
fmt dd72a59
add tea.yaml 70e4c1b
https://github.com/juliangruber/brace-expansion/compare/v3.0.0.....
OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting @dependabot ignore this major version or @dependabot ignore this minor version.
If you change your mind, just re-open this PR and I'll resolve any conflicts on it.
DONT LET THE HATERS BRING YOU DOWN DEPENDABOT
echo run-shortcut:pushToTalkNormalStart | nc -U $XDG_RUNTIME_DIR/vesktop.sock
arch users: pacman -S nmap, then use ncat instead of nc
how do you use ts im stupid and js want my deafen bind
git clone -b global-shortcuts https://github.com/Vencord/Vesktop.git
cd Vesktop
pnpm start
add a keybinding to echo run-shortcut:toggleMute | nc -U $XDG_RUNTIME_DIR/vesktop.sock (the exact way depends on your desktop environment, you'll have to look that up yourself)
Flatpak Screenshare Fix — Configuration Side
In addition to the code fixes in this PR, there is a Flatpak configuration issue that independently breaks screen
sharing on Wayland. Documenting here for anyone hitting screenshare problems on the Flatpak build.
Problem
The Flatpak wrapper script startvesktop does not read ~/.var/app/dev.vencord.Vesktop/config/electron-flags.conf.
Any Chromium flags placed in that file (e.g. WebRTCPipeWireCapturer) are silently ignored. ...
that is also complete nonsense, please just go away with your AI
Content
your a actual fucking retard you know that spam a cause ai your a literately retard the discord screensharing is a issue cause ecltron and the picker causes a session to expire because of how ecltron works you litterly get cancer and actually die im trying to help you and you act retarded i apy for this its not 2021 !!!!!!!!!!!!!!!!!!!!
im gunna relaese zero days the whole 2 months and make working poc just to prove a point so all your suers can be hacked
chuddy gonna zeroday discord
least insane vencord repo email
Symptom
In a flatpak install, toggling "Start With System" succeeds and writes ~/.config/autostart/dev.vencord.Vesktop.desktop, but the autostart unit crashes Chromium on next login:
FATAL:sandbox/linux/suid/client/setuid_sandbox_host.cc:166
The SUID sandbox helper binary was found, but is not configured correctly.
Rather than run without sandboxing I'm aborting now. You need to make sure
that /app/bin/vesktop/chrome-sandbox is owned by root and has mode 4755.
`systemd-xdg-aut...
Perfect example of someone (immature) blindly trusting anything ChatGPT says
Perhaps I should give a status update. I'm struggling to reliably detect when the screen is locked in my DE that is COSMIC.
org.freedesktop.ScreenSaver.GetActive API is a nonstandard extension implemented at least by KDE: https://github.com/pop-os/cosmic-idle/issues/5
Based on a quick query through the code it's not implemented by Sway/wlroots, Niri, Jay and Hyprland.
I've also tried to listen on org.freedesktop.login1.Session interface but on COSMIC only lock events appear there: ...
Oh I completely forgot I had planned to work on this over the weekend T-T. Maybe I do have commitment issues. Anyways, as far as I’m aware it’s by design that lockscreens can’t be detected on wayland so I’m still not sure if it’s something worth pursuing. If you do get it to work that’d be great though! I’m also curious how you’re implementing the lockscreen detection into the powerMonitor api if you’ve done anything on that front yet. I want to slightly tweak the current isIdle api but I kee...
Can confirm that the same issue exists on Windows. Patching the install fixed the issue immediately.
got hit by this bug 2 days in a row oh my god
@MayasMischief Have you tried right-clicking the tray icon and selecting "Repair Vencord"?
Hey, void linux user here. I have had this issue on my hardware as well and I have found that after extensive testing, only the rpm official package works properly, which also appears to be what the vesktop-bin AUR package sources.
I have had no luck with the flatpak, tarball, deb package, or building from source on multiple different systems (only amd gpus and igpus seem to be affected from my tests). Simply unpacking the rpm and running the binary out of opt however works great even ...
screen sharing not working on Plasma.
I'm using the AUR package of Vesktop on CachyOS 7.0.9 and screensharing technically "works" for me on KDE Plasma 6.6.5, but the stream very often corrupts and appears extremely distorted:
<img width="1401" height="708" alt="Corrupted stream" src="https://github.com/user-attachments/assets/b015c27b-865a-4f9f-b8b0-fe70b46c9fa3" />
This is a direct screenshot of what the source looks like without any corruption:
<img width="1920" height="1080" alt="Scre...
What I mainly need is testing on different desktop environments, so you can help with this and report back.
Just thought I'd add here that I was able to successfully get this working for me
Built the global shortcuts branch and tested on my machine (CachyOS + KDE Plasma)
Went into Settings in Vesktop and added keybind there. (Accepted Global shortcut requested prompt)
Then had to go into KDE Settings > Keyboard > Shortcuts, and add keybind there to match with what I entered insid...
I've this exact same issue and it's definitely vesktop specfic because the normal client works perfectly fine
What I mainly need is testing on different desktop environments,
Working on COSMIC:
<img width="1222" height="747" alt="image" src="https://github.com/user-attachments/assets/95c36238-c5b6-4fa3-97b5-eb653a829017" />
And on KDE:
(i'll edit this comment after i swich DEs)
Having tested again on KDE Plasma after https://github.com/Vencord/Vesktop/pull/1238#issuecomment-4421263955, I double down on the issue coming from most probably using duplicated states for the keybinds instead of vesktop being stateless for them and relying as much as possible on the native keybind states
Also, it's not possible to register a keybind from KDE Plasma's keybind settings, which makes me think they have not been declared and registered properly and are currently implemented as imperatively asking for permissions with a persisting state in vesktop
Instead, they should be declared in some way so that the permission prompt is not required, and to make it possible to register the keybind from the DE settings directly
The state then becomes the DE instead of in Vesktop
I guess...
ah nice, i didn't know the portal "standardization" was still garbage yet, thks
no its literally that the impl in that pr is bad
wayland global shortcuts are completely different from every other platform
you register them all at launch
you cant register shortcuts one at a time
I mean you could technically create a new session without killing the app to do this but 💀
then we would have to hardcode keybinds and hope the desktop supports changing it and users are smart enough
the pr impl works fine on kde
which is the only backend that has a sane impl
the portal is just atrocious and not user friendly
also if I installed an app and immediately got prompted to add global keybinds on first launch i would uninstall it again
hella annoying
its in the spec that you can just not declare defaults (also just dont use a bad desktop)
cope just press cancel
nope
it doesnt have to register right at launch it can be when the user navigates to the keybind UI or whatever
and the app can request through the portal to open the keybinds settings
all your issues are from the gnome experience and being used to apps owning global keybinds
not with electron api
not with electron api
yeah cause electron is bad
yes and this is what 95% of people are used to
vee pay me money and ill finish my pr that makes electron global shortcuts actually work properly
I dont understand why wayland insists on doing everything differently
it just hurts user friendliness
security or whatever
it will never be the year of the linux desktop with the bullshit wayland is doing
security is good but not when it conflicts with user friendliness, at least not in something like a personal computer 
honestly I like waylands shortcut impl cause then all the configurations are in one place
99% of people dont care
yeah but most desktops don't even support it and no push to talk 😭
the user friendliness is fine if the app dev does Ir correctly
wayland global shortcuts has up down events electron/chromium is just bad
. (ill kms before I touch electron again)
genuinely insane that u need to compile and download all of chromium to make like a two like change
seven billion gigs
build an electron alternative @humble mortar
you cant say tauri etc they're even worse
actually horror
I dont even think tauri has a wayland global shortcuts impl
and the fact that they rely on gtkwebview is so terrible
Also, push to mute is usful when dealing with a in-game voicechat and being in a unrelated discord voicechat
I also am looking for push to mute, not just toggle mute or push to talk, for this exact reason. Open mic in Discord, and momentarily mute Discord throughout the duration of talking in-game. For the official client, my push to mute keybind in Discord is the same keybind as push to talk in games.
I had the same problem and it was caused by having notification daemon not turned on. If you're experiencing this problem turn on your notification daemon.
Bazzite user, same issue reproduced but it's inconsistent. I'll have times where restarting the client or just trying again fixes it. Very random,.
the global keybinds pr is similar to the thing i did to get global keybinds to work
just adding a socket and using something else to send messages to it
thats amazing lmao
stupid question / probably been considered already but if electron's implementation of keybinds sucks so bad then why not make some lightweight process to run alongside vesktop and handle all the keybinds itself?
tuxinal made a pr that did just that but it was kinda mid and smth else I forgor what exactly was the other major issue but adding something to libvesktop instead of a new thing would probably be worth considering given it already deals with dbus malware
it supports keybinds I'm not sure if it supports global correctly instead of being a keylogger
it's a keylogger
not on wayland i think
or if it does its very new
but doubt they do
I just tested they don't
I really don't want anything new to maintin
lmao
one of the main reasons i rejected the pr
the responsibility would still be mine 
I might try some bs tomorrow just to see how much effort it'd be to add to libvesktop
Content
you can technically disable it by disabling aarpc entirely but thats not pratical
I have been using a patched version of vesktop with this PR for a while now and it works perfectly
I second this. Works without any issue when switching on/off multiple different VPNs.
does vesktop run its own electron?
nop
@MayasMischief Have you tried right-clicking the tray icon and selecting "Repair Vencord"?
I've tried this repeatedly with no success, this bug seems to be recurring.
tbh this is only thing i missing in vesktop
Just tested this PR in Fedora 44 with KDE Plasma 6.6.
It does trigger the global shortcut portal but only after I closed and relaunched the app, and only for a single shortcut (the other one seems to be ignored)
I got all of my shortcuts to be registered. But removing some of shortcut on Vesktop wouldn't remove them from system shortcuts as @Malix-Labs stated before.
<img width="1929" height="450" alt="изображение" src="https://github.com/user-attachments/assets/91731c48-43e6-435f...
Please look into merging this, works perfectly, and it's super annoying having to turn off tailscale, join a voice channel, then turning tailscale back on when dealing with SSH
