Help with YT-Cipher

[slim current]

The this content isnt available is just that, youtube is saying "this video is not available" could be many reasons, geoblocked, client restricted, etc. not cipher related most likely

[eager pewter]

I tried turning off oauth then it says this video requires login then with oauth enabled saying this content isnt available

[slim current]

unfortunately YT wont tell us why. You can always open another help thread about it or search for that error message on here, I am sure there are more threads about it

[eager pewter]

ok

[wild aspen]

His oauth account got rate limited (a single account can request only ~300 /player requests per hour )

[eager pewter]

Same issue with another refresh token and another hosting ip

[slim current]

Its possible that video is just unavailable to tv clients or smth, its for just one video right?

[wild aspen]

some music videos don't work with non-music clients

[still spoke]

@slim current Do you know why it does that?

[slim current]

com.sedmelluq.discord.lavaplayer.source.youtube.YoutubeSignatureCipherManager.extractFromScript means the youtube-source plugin is not active, refer to #1421172218339000491 message
As well as the https://github.com/lavalink-devs/youtube-source?tab=readme-ov-file#plugin documentation

[still spoke]

I already have all of that and I disabled YouTube

[crude nymph]

Speak english pls, read rule

[still spoke]

You have to come and comment on your stupidity 🤔 I'm speaking with a German I didn't pay attention to the translator why are you pinging me

[scenic fox]

Do you want a timeout lol

[scenic fox]

The only stupidity is this comment right here

[crude nymph]

"stupidity" ?? I'm just telling you to speak English simply, that's all, so avoid being aggressive.

[still spoke]

What about you, the only thing you do that can be unpleasant all the time?

[still spoke]

I'm not being aggressive, just being realistic. You ping him to tell me, rule, yes, well, I didn't notice that the translator was still on German, know everything.

[scenic fox]

feel free to leave this server if you dont like it

[slim current]

Either you do not, or your configs are not being applied properly or something. We cannot fix this for you, we can only point you to the documentation to help yourself. I recommend taking a close double and triple checks of your configs and the errors. You have demonstrated several times now that you do not read docs, error messages, etc. You just see it and ask "Fix it for me pls". We are more than willing to help, but it really feels like this is going in circles and you should really try to learn and figure it out. I garuntee that will be a much better outcome than asking everyone else to fix it for you and hoping you get lucky. If you want to keep trying that though, try asking one of the AIs for help. I feel like I have given you all the help I can

[slim gate]

do you want me to pin this

If you wanna make an FAQ at any point, just tag me and I'll pin that too

[still spoke]

Very good so I would ask for more help

[patent cargo]

Hello

[eager pewter]

#general

[slim current]

YT-cipher FAQ

--- What is the easiest way to set this up? ---

1. Make sure you are on the latest release of youtube-source plugin
2. Add this to your lavalink application.yml in the section under plugins -> youtube

plugins:
  youtube:
    remoteCipher:
      url: "https://cipher.kikkia.dev/"

*** --- Why do I need yt-cipher ---***
YouTube constantly changes a hidden code needed to play videos. Instead of having Lavalink break this code repeatedly, we made a separate, smaller service to handle just that one job.

This makes it much faster and easier for the Lavalink team to fix things every time YouTube makes a change.

***--- What if I want to host the yt-cipher service myself? --- ***
It is pretty easy to setup yt-cipher yourself, we reccomend using docker as that is the easiest. Please see this section of the docs: https://github.com/kikkia/yt-cipher?tab=readme-ov-file#hosting-yourself

*** --- I think yt-cipher is causing an error what should I do? ---***

Look at the error, if search the error and see anything referencing RemoteCipherManager definitely reach out. If you see 403 errors, also reach out.

--- This video is not available or Please sign in
These are likely not yt-cipher, and I recommend opening a general help thread.

--- Problematic YouTube player script or LocalSignatureCipherManager
This is likely an issue with your remoteCipher setup in your lavalink application.yaml, please double check spelling and indentation. It should be nested in plugins -> youtube. Also double check that lavalink -> server -> sources -> youtube == false

*** --- yt-cipher doesnt work it 404s. --- ***
This is by design, all paths will 404, except for the paths defined in the api specification: https://github.com/kikkia/yt-cipher?tab=readme-ov-file#api-specification

[slim current]

We can pin this, also maybe a fresh help thread just called YT-Cipher FAQ, maybe or if possible renaming this thread

[crude nymph]

Hello @slim current, I saw your UserAgent idea, which I think is quite good. I find it interesting, but I know it's still in development. I suppose it can be customised on the YouTube source side or on the Yt-Cipher side ?
I saw that it can remain optional. I find it interesting and intriguing to learn more about this idea and how it will be used if it is customisable or something left as default.

[slim current]

Its nothing too fancy, its only used for prometheus metrics.
For example, I run several LL nodes, so I give each node a unique user-agent, then I can see in my metrics dashboards stuff like requests per second, what player scripts are being used for each node, etc

Another use for me, is with the public one, I can now track users YT-source version number, so if we have to make breaking changes to api spec or something we can have a better idea of how that effects users

[slim current]

Going to do a small dns update, I dont expect more than a few seconds of downtime, but I thougt I should post this here to get ahead of it in case a few people hit some failed requests

Ok CF Dns is shitty rn, if you swap to ctest.kikkia.dev it will work.

Ok fixed ~10 mins of downtime, sorry

[slim current]

With that migration finished though, it should shave a few ms off of response time. also cipher.kikkia.dev/ and cipher.kikkia.dev/api/ both work. And if you use the one without /api/ that should shave another 1-2ms off on average.

[burnt token]

2025-10-07T11:15:37.463Z  INFO 32 --- [Lavalink] [  XNIO-1 task-2] l.server.player.AudioLoaderRestHandler   : Loaded track ROSÉ & Bruno Mars - APT. (Official Music Video)
2025-10-07T11:15:37.507Z  INFO 32 --- [Lavalink] [  XNIO-1 task-2] l.server.io.RequestLoggingFilter         : GET /v4/loadtracks?identifier=https://www.youtube.com/watch?v%3Dekr2nIex040, client=103.183.82.227
2025-10-07T11:15:38.036Z  INFO 32 --- [Lavalink] [  XNIO-1 task-3] l.server.io.RequestLoggingFilter         : PATCH /v4/sessions/x894xtt96koi0frq/players/904230958465699870, client=103.183.82.227, payload={"voice": {"token": "34a77e6c03589e1c", "endpoint": "india10010.discord.media:443", "sessionId": "8f8a1389d8cbf3f66ce4935c21dcbc67"}}
2025-10-07T11:15:38.258Z  INFO 32 --- [Lavalink] [  XNIO-1 task-3] l.server.io.RequestLoggingFilter         : PATCH /v4/sessions/x894xtt96koi0frq/players/904230958465699870?noReplace=False, client=103.183.82.227, payload={"encodedTrack": "QAAA0wMAMVJPU0XMgSAmIEJydW5vIE1hcnMgLSBBUFQuIChPZmZpY2lhbCBNdXNpYyBWaWRlbykABVJPU8OJAAAAAAACo8gAC2VrcjJuSWV4MDQwAAEAK2h0dHBzOi8vd3d3LnlvdXR1YmUuY29tL3dhdGNoP3Y9ZWtyMm5JZXgwNDABAD9odHRwczovL2kueXRpbWcuY29tL3ZpL2VrcjJuSWV4MDQwL21heHJlc2RlZmF1bHQuanBnP3Y9NjcxMWRiYWMAAAd5b3V0dWJlAAAAAAAAAAA=", "position": "0"}
2025-10-07T11:15:38.292Z  INFO 32 --- [Lavalink] [back-1-thread-1] d.l.y.http.YoutubeAccessTokenTracker     : Updating YouTube visitor id succeeded, new one is CgtMRUVaZllVSEhybyja7JPHBjInCgJGUhIhEh0SGwsMDg8QERITFBUWFxgZGhscHR4fICEiIyQlJiAXOgwIASD_pdToocu98mhYlsn508Og9800, next update will be after 600 seconds.
2025-10-07T11:15:39.492Z  WARN 32 --- [Lavalink] [back-1-thread-1] c.s.d.l.t.p.LocalAudioTrackExecutor      : Suspicious exception for playback of ekr2nIex040
com.sedmelluq.discord.lavaplayer.tools.FriendlyException: This video is unavailable
        at dev.lavalink.youtube.track.YoutubeAudioTrack.process(YoutubeAudioTrack.java:131) ~[youtube-plugin-1.14.0.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:109) ~[lavaplayer-2.2.3.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.player.DefaultAudioPlayerManager.lambda$executeTrack$2(DefaultAudioPlayerManager.java:339) ~[lavaplayer-2.2.3.jar!/:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:840) ~[na:na]

Do you have any idea on wht the issue is

[crude nymph]

Enable Oauth and Use clients

Or the video is now private or simply no longer available.

[slim current]

For that address error @sand kite What is the host set to?

[sand kite]

i set host to my host ip but it doesn't seems to work so i changed server.ts that || part "0.0.0.0" to my host's ip

[slim current]

Is this within a docker container?

[sand kite]

wait to clarify i'm using a aio server to run deno and also 2 days before i changed my server's region so before that it worked correctly but after that its giving me this error.

[slim current]

If you just leave it at 0.0.0.0 it will listen on all network interfaces, the error you are seeing is due to deno not being able to find a network interface for the address you have set

[sand kite]

oh

so i need to give it like 0.0.0.0:1646

like this?

i mean port and ip

[slim current]

Just 0.0.0.0

Port is separate

[sand kite]

yeah

i mean the port i need to give that right?

[slim current]

You can use any port you want
Then with host set to 0.0.0.0 the service is accessable at http://your-ip:your-port

These can be set as env variables, or leave them at default

PORT - Port to run the api on, default: 8001
HOST - Sets the hostname for the deno server, default: 0.0.0.0

[sand kite]

this not working

i gave ip and port but it uses default

[slim current]

How are you running the service, its likely something is overwriting your env variables

[sand kite]

as i said i'm using the aio server

in Pterodactyl panel

[slim current]

oh god pterodactyl

[sand kite]

Yes.

[slim current]

Umm there is some other stuff from a while back with other users and pterodactyl over writes their env variables

I think they found a fix, let me see if I can dig it back up

[sand kite]

Yeah same for me also

but i choose editing server.ts :)

btw what is this error??

2025-10-08T11:40:41.568-04:00 ERROR 31 --- [Lavalink] [ce-token-poller] d.l.youtube.http.YoutubeOauth2Handler    : Failed to fetch OAuth2 token response
java.io.IOException: Invalid status code for oauth2 token fetch: 400
        at com.sedmelluq.discord.lavaplayer.tools.io.HttpClientTools.assertSuccessWithContent(HttpClientTools.java:168) ~[lavaplayer-2.2.3.jar!/:na]
        at dev.lavalink.youtube.http.YoutubeOauth2Handler.pollForToken(YoutubeOauth2Handler.java:155) ~[na:na]
        at dev.lavalink.youtube.http.YoutubeOauth2Handler.lambda$initializeAccessToken$0(YoutubeOauth2Handler.java:108) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:1583) ~[na:na]

[slim current]

Just change these to what you want

const port = Deno.env.get("PORT") || 8001;
const host = Deno.env.get("HOST") || '0.0.0.0';

[sand kite]

yeah i already done.

[sand kite]

this error is now spamming in console.

[slim current]

That appears to be an issue with the youtube-source oauth flow. Might need to reset your refresh token?

Nothing to do with yt-cipher tho

[sand kite]

Ah i don't use refreshToken because its not working :)

anyways thanks for the assistance now server is working :)

btw kikkia can u come on my dm?

topi sir just asking a doubt nothing else :)

[turbid loom]

does .env files not work for those btw?

[slim current]

https://tryitands.ee/

I don't see why not, but I only ever run this in docker so

[turbid loom]

Hm, I will give it a shot and come back here if it does or not

well it does work :D

just that u forgot to mention --env in the readme, and only had it set to deno run --allow-net --allow-read --allow-write --allow-env which for some reason doesn't take env variables without the --env flag included as well, extra steps are annoying 💀

[hoary dust]

look like yt-cipher uses a lot ram

[turbid loom]

yeah but it's not that much tbh

[hoary dust]

Looking at my log it seems like it's being used quite a bit but it's got a lot of RAM

[violet flame]

Idk why I get:
org.apache.http.client.ClientProtocolException: URI does not specify a valid host name: http://0.0.0.0:8001/
On lavalink it has the correct ip and port, the cypher works like it should, gives the script but fails here, how can i fix it?

[scenic fox]

0.0.0.0 is not a valid ip

[turbid loom]

127.0.0.1 should work :kek

[violet flame]

On lavalink it is

Alright?

[scenic fox]

no

0.0.0.0 is NEVER a valid ip

[crude nymph]

0.0.0.0 is work

for me

[slim current]

I have it on my backlog to see if we can make deno use less. The cipher stuff is decently memory heavy, especially caching the player scripts. It does "feel" high. Its possible some of the ast libraries hog some. Also iirc deno doesnt garbage collect till its reaching near the ram limit.

At some point if i have a lot of time i also want to slim the docker image size

[slim current]

0.0.0.0 works as a hostname for deno, but is not a valid ip for a client to make a request to.

0.0.0.0 isnt actually an ip when used as a hostname, its just saying "attach to all network interfaces "

[eager pewter]

is there any issue with the latest yt plugin v1.14.0 because that is not working for me even with cipher but at the same time this snapshot with same config and cipher is working.

dev.lavalink.youtube:youtube-plugin:755aab35029153a705ebd4be63c56a55d68ae678

[slim current]

"not working for me", need more description. Errors? Logs? Etc

[slim current]

Config naming changed a little, double check your yaml vs the current github documentation

[eager pewter]

working with this structure

remoteCipher:
         url: ...
         password: ...
         userAgent: ...

and the snapshot was working with this

remoteCipherUrl: ...

[silver salmon]

btw

2025-10-09T12:34:31.120Z ERROR 36 --- [Lavalink] [back-1-thread-1] d.l.y.c.LocalSignatureCipherManager : Problematic YouTube player script /s/player/38e0f4b0/player_ias.vflset/en_US/base.js detected (issue detected with script: must find sig function). Dumped to /tmp/lavaplayer-yt-player-script882335274010173712.js (Source version: 755aab35029153a705ebd4be63c56a55d68ae678)

new player script

[eager pewter]

it's working after I changed to new yml format

[silver salmon]

yep

once i go back to 1.14.0 it works fine

💀

[slim current]

They change player scripts pretty much every day, except on weekends

[eager pewter]

YouTube devs are employed, but every day they prove why they shouldn’t be.

[violet flame]

Bruh

It’s for not sending my ip

On lavalink is correct

I’m not gonna send the ip of my server hosting cypher

It gets the script but it fails on that

[scenic fox]

Then don't write 0 lol

[violet flame]

Someone told me to .-.

[scenic fox]

Use x xd

[violet flame]

Alright

Let me send the message correctly .-.

Idk why I get:
org.apache.http.client.ClientProtocolException: URI does not specify a valid host name: http://x.x.x.x:8001/
On lavalink it has the correct ip and port, the cypher works like it should, gives the script but fails here, how can i fix it?

Now better?

[scenic fox]

And your config looks like?

[wild aspen]

your lavalink container and yt-cipher should use the same network otherwise u will get this error

[violet flame]

Why so? It’s doing an http request and when I do it manually it works

Only fails when lavalink does it

[violet flame]

Looks like the example, it’s a copy paste but give me a sec I get home to paste it

[scenic fox]

Manually from where

[wild aspen]

ur using http://ip:8001 i thought u have another container with this name 💀

[violet flame]

When I do the request using things like postman to test why it fails I don’t understand what’s wrong on that endpoint

Because it does work correctly the server

So I don’t see from where is getting that is not a valid host

[violet flame]

My config.yml

[slim current]

url: http:///x.x.x.x:8001
Wanna double check that the third / isn't causing this issue?

[violet flame]

It's not, as the script is working... Check the error log you'll see

[violet flame]

.

[scenic fox]

open ya eyes @violet flame

/ / /

xd

add quotes

[slim current]

https://ideone.com/EjXoLJ
Three slash parses the host as null, therefor no valid host name error

[violet flame]

I don't think it's that, but let me check anyways .-.

[violet flame]

But why it prints it with two then

And it does fetch the script

[slim current]

How does it fetch the script if its throwing that error?

That would go before the http request fires since its saying it doesnt have a hostname to send it to

If that doesn't fix it, post a full stacktrace

[violet flame]

Bruh, it was something else, the fucking VPS company changed my IP 💀

Though the log doesn't actually make sense

[scenic fox]

XD?

Dude
A triple slash isn't valid either way

[violet flame]

2025-10-09T06:42:23.663Z DEBUG 13 --- [Lavalink] [XNIO-1 task-4] d.l.youtube.cipher.RemoteCipherManager   : Timestamp from script /s/player/38e0f4b0/player_ias.vflset/en_US/base.js
2025-10-09T06:42:23.664Z DEBUG 13 --- [Lavalink] [XNIO-1 task-4] d.l.youtube.cipher.RemoteCipherManager   : Getting timestamp for script: /s/player/38e0f4b0/player_ias.vflset/en_US/base.js
2025-10-09T06:42:23.664Z DEBUG 13 --- [Lavalink] [XNIO-1 task-4] d.l.youtube.YoutubeAudioSourceManager    : Client "WEB" threw a non-fatal exception, storing and proceeding...

org.apache.http.client.ClientProtocolException: URI does not specify a valid host name: http://x.x.x.x:8001/get_sts```

[violet flame]

That was by accident I guess when I was censoring everything

As when I checked the server's config there were two

[scenic fox]

[violet flame]

Either way, on the new ip it works

Thanks for helping me on this... Garbage, let's leave it there

[floral island]

why is it asking me an api token on localhost?

Caused by: java.io.IOException: Decryption proxy request failed with status code: 401. Response: {"error":"Missing API token"}

with "https://cipher.kikkia.dev/" works fine

remoteCipher:
      url: "http://0.0.0.0:8001"
# Already tested http://localhost:8001, same api token error

obs: im using docker

Fix:

[turbid loom]

Because it's self hosted?

And the other is public without a password set

[floral island]

good point

[nova bridge]

can i please use somebody elses docker image?

[turbid loom]

Just run it with deno directly

[nova bridge]

?

i pay to deploy apps to railway.com and would like a docker image. never heard of deno

[turbid loom]

deno is a javascript runtime

https://github.com/kikkia/yt-cipher moreover, the documentation is in the readme

[nova bridge]

services:
  ejs-api:
    # Uncomment to build locally
    # build:
    #   context: .
    #   dockerfile: Dockerfile
    image: shmuelley/yt-cipher:latest
    container_name: ejs_signature_api
    ports:
      - "8001:8001"
    environment:
      - API_TOKEN=noctra
    restart: unless-stopped

?

im getting errors

2025-10-11T06:10:38.303Z ERROR 26 --- [Lavalink] [back-1-thread-1] c.s.d.l.t.p.LocalAudioTrackExecutor      : Error in playback of 5rqJJfabAujtHaGuZCXy4q
com.sedmelluq.discord.lavaplayer.tools.FriendlyException: Something broke when playing the track.
        at com.sedmelluq.discord.lavaplayer.tools.ExceptionTools.wrapUnfriendlyExceptions(ExceptionTools.java:45) ~[lavaplayer-2.2.3.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:121) ~[lavaplayer-2.2.3.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.player.DefaultAudioPlayerManager.lambda$executeTrack$2(DefaultAudioPlayerManager.java:339) ~[lavaplayer-2.2.3.jar!/:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:642) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:1583) ~[na:na]
Caused by: java.io.IOException: Decryption proxy request failed with status code: 301. Response: 
        at dev.lavalink.youtube.cipher.RemoteCipherManager.getTimestampFromScript(RemoteCipherManager.java:270) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.cipher.RemoteCipherManager.getTimestamp(RemoteCipherManager.java:127) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.clients.skeleton.NonMusicClient.loadTrackInfoFromInnertube(NonMusicClient.java:130) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.clients.skeleton.StreamingNonMusicClient.loadFormats(StreamingNonMusicClient.java:35) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.loadBestFormatWithUrl(YoutubeAudioTrack.java:211) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.processWithClient(YoutubeAudioTrack.java:148) ~[youtube-plugin-1.14.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.process(YoutubeAudioTrack.java:94) ~[youtube-plugin-1.14.0.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.DelegatedAudioTrack.processDelegate(DelegatedAudioTrack.java:25) ~[lavaplayer-2.2.3.jar!/:na]
        at com.github.topi314.lavasrc.mirror.MirroringAudioTrack.process(MirroringAudioTrack.java:56) ~[lavasrc-plugin-4.8.1.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:109) ~[lavaplayer-2.2.3.jar!/:na]
        ... 4 common frames omitted

[slim current]

You can always use the public api https://cipher.kikkia.dev/ or use the official docker image.
There is nowhere in the service that returns any 300 codes, so that is likely from whatever you are hosting it on

[turbid loom]

I was just about to say it's caused by railway itself

People should really stop using such stupid things and move to a vps

[slim current]

If the actual response still makes it to the client, just with a 300 code maybe we can relax the plugin to treat it as a non error, but idk how these hosts handle their http gateways and what not

[wet sundial]

@slim current hey bro the read timeout error most of the time any fix of this ?

[silver salmon]

lavalink:
  server:
    timeouts:
      connectTimeoutMs: 10000
      connectionRequestTimeoutMs: 10000
      socketTimeoutMs: 10000```

add timeouts

[wet sundial]

i use this still

[silver salmon]

then up the timeouts

[wet sundial]

Did you tell me that if I use these timeouts, any logs will show in LL to confirm it's working?

[scenic fox]

Send your config
Then we can tell you if it's correct or not

But if 10s isn't enough timeout you should consider if something else is broken in your setup

If it doesn't work, could also be that your lavalink is outdated

[slim current]

Also if you are on a really old yt-cipher and a slow machine the AST on script download could hit the 3s timeout by default.

[wet sundial]

no i updated it
i use hetzner.

[scenic fox]

Show your logs then
It shows all the versions

full logs

[slim current]

Also, i have probably asked you one of the most important questions half a dozen times to no answer, how often does this happen? 100% of the time? 10%? <1%?

[wet sundial]

here is my config

[wet sundial]

no not 100% of time

sometime

[slim current]

So this happens say after X video loads, with the same player script that the others used?

Just making sure, because if this happens on a script thats already been processed and cached its very likely some networking hangup on the machine or smth

[wet sundial]

am using ipv6 is it for that ?

or i think sometime network hangup hope so

[slim current]

How long does it take on a fresh restart to happen? Minutes, hours? If you had to guess at how many videos are loaded before it happens how many do you think?

Couple things to try:

• Setup a curl overnight or something to hit it like every couple seconds and see if those ever timeout (if possible over ipv4 and ipv6 separatetly to see if it has any effect)
• Try hitting the public api for a while and see if those ever time out

Do you run in docker?

[wet sundial]

yes docker

[slim current]

Maybe we could have you add a debug log that just prints out the request duration within the yt-cipher server, so we can see if any requests actually take long enough to trigger a timeout

So we can determine if its yt-cipher being super slow for some reason or if its some networking thing somewhere

[wet sundial]

okay

[slim current]

If you add a try / finally block around all of the baseHandler function, and log the total duration elapsed in the finally block that should log it on every request.

In the docker-compose file uncomment out the build, context and dockerfile lines, and make the image line commented.

Then just run docker-compose build to build a fresh image with your new logs

[nova bridge]

what is required in the LL youtube config to play a track?

[slim current]

https://github.com/lavalink-devs/youtube-source?tab=readme-ov-file#lavalink-2

url = your self hosted instance or the public instance https://cipher.kikkia.dev

[nova bridge]

and run on 1.14?

ok thank you. im getting a load failed now but ill figure that out

[nova bridge]

lavalink wants me to login. how does that work?

[silver salmon]

use oauth

[nova bridge]

holy shit... fixing my bot was quite the adventure. thanks 🙏

u too kikka

i got my cipher to work too. i had http instead of https

[sand kite]

Wait what is the use of reolve_url??

[silver salmon]

https://github.com/kikkia/yt-cipher?tab=readme-ov-file#post-resolve_url

tells you

[turbid loom]

But wait, is it automatically picked up by the yt plugin?

or nah?

[meager moon]

Having the same issue

[slim current]

Its a new endpoint that does a lot more work than decrypting the cipher, it also constructs the full playback url
Make sure to update your yt-cipher service (if you self host) before doing the update of the plugin

Caused by: java.io.IOException: Decryption proxy request failed with status code: 301. Response:?
This means that your host you are running yt-cipher on is messing with the http responses from yt-cipher, you may need to mess around with it to figure out how to make it let you talk directly to yt-cipher

[nova bridge]

i can help you if you want

[meager moon]

Yes Please

[nova bridge]

want to start a call or move to a vc?

[meager moon]

Is it possible to chat here?

[nova bridge]

yeah

[meager moon]

currently i m out of home, So, I couldn't attend calls rn

Wait let me share my app.yml

@nova bridge

My app.yml

[nova bridge]

what are you hosting the lavalink server on?

locally?

[meager moon]

my vps

[nova bridge]

so you probably wouldnt use local host right? they are deployed separate?

[meager moon]

no

[turbid loom]

you would if both are hosted on the same vps

[nova bridge]

right

[meager moon]

yt cipher , my bot and lavalink are hosted on same vps

[nova bridge]

Here is my app.yml for reference. my lavalink and cipher run separate.

lavalink:
    plugins:
        - dependency: dev.lavalink.youtube:youtube-plugin:1.14.0
          snapshot: false
        - dependency: com.github.topi314.lavasearch:lavasearch-plugin:1.0.0
          snapshot: false
        - dependency: com.github.topi314.lavasrc:lavasrc-plugin:4.8.1
          snapshot: false
    server:
        bufferDurationMs: 400
        filters:
            channelMix: true
            distortion: true
            equalizer: true
            karaoke: true
            lowPass: true
            rotation: true
            timescale: true
            tremolo: true
            vibrato: true
            volume: true
        frameBufferDurationMs: 5000
        gc-warnings: true
        nonAllocatingFrameBuffer: false
        opusEncodingQuality: 10
        password: youshallnotpass
        playerUpdateInterval: 5
        resamplingQuality: LOW
        soundcloudSearchEnabled: true
        sources:
            bandcamp: true
            http: true
            local: false
            nico: true
            soundcloud: true
            twitch: true
            vimeo: true
            youtube: false
        timeouts:
            connectTimeoutMs: 10000
            connectionRequestTimeoutMs: 10000
            socketTimeoutMs: 10000
        trackStuckThresholdMs: 10000
        useSeekGhosting: true
        youtubePlaylistLoadLimit: 6
        youtubeSearchEnabled: true
logging:
    file:
        path: ./logs/
    level:
        lavalink: INFO
        root: INFO
    logback:
        rollingpolicy:
            max-file-size: 1GB
            max-history: 30
    request:
        enabled: true
        includeClientInfo: true
        includeHeaders: false
        includePayload: true
        includeQueryString: true
        maxPayloadLength: 10000
metrics:
    prometheus:
        enabled: false
        endpoint: /metrics
plugins:
    lavalyrics:
        sources:
            - spotify
            - youtube
    lavasrc:
        applemusic:
            albumLoadLimit: 6
            countryCode: US
            playlistLoadLimit: 6
        providers:
            - ytsearch:"%ISRC%"
            - ytsearch:%QUERY%
            - scsearch:%QUERY%
            - spsearch:%QUERY%
            - sprec:%QUERY%
        sources:
            applemusic: true
            deezer: false
            flowerytts: false
            spotify: true
            vkmusic: false
            yandexmusic: false
            youtube: true
        spotify:
            albumLoadLimit: 6
            clientId: ###
            clientSecret: ###
            localFiles: false
            playlistLoadLimit: 6
            preferAnonymousToken: true
            resolveArtistsInSearch: false
    youtube:
        clients:
            - TV
            - TVHTML5EMBEDDED
        oauth:
            enabled: true
        remoteCipher:
            password: noctra
            url: https://cipher-socket-production.up.railway.app
sentry:
    dsn: ""
    environment: ""
server:
    address: 0.0.0.0
    http2:
        enabled: false
    port: 31663

[meager moon]

So, You have hosted your yt cipher on railway?

[nova bridge]

you dont need any other youtube settings besides what i have set and as far as i am aware, the 301 status is a problem with the cipher.

yes and you may use it if you would like

[turbid loom]

Cipher works fine for me 🤔

and i also use vps

[meager moon]

Well i am trying to Deploy but it showing this

[slim current]

301 status code is never returned by cipher. In schmuelly's case railway is doing something in between his LL and yt-cipher, and edits the response code or something.

Post your full stacktrace @meager moon, If you are not on railway i doubt its the same error

[nova bridge]

i wouldnt just deploy to railway because i pay to keep my services online.

[meager moon]

I am Getting the error that Something Broke When Playing Track....

And Missing Api ... in Yt cipher , but i have already hosted it properly

Using docker

[turbid loom]

that's a token issue

your token might be invalid in the docker compose file

[meager moon]

Not Possible i had checked my yt cipher .env

[turbid loom]

and the app.yml

[turbid loom]

hm

[meager moon]

I have double checked its correct

[slim current]

Not the full error, but my guess that Missing Api ... is missing api token?
Also the Something Broke When Playing Track is not the actual error, read the whole stack trace you will find caused by: ACTUAL ERROR

[turbid loom]

docker doesn't take .env if i remember right

or does it?

[meager moon]

Let me share.. my logs

[slim current]

It does if you volume it in I think

[turbid loom]

i run it with deno directly

because with docker it just caused issues whenever i modified the token in composer file

💀

it instantly crashed

and worked fien with the defaults

fine*

[slim current]

With docker it only has access to the .env file if you specifically add a docker volume in the compose so that it can actaully access the .env file on your fs

[turbid loom]

Aha

[slim current]

But just using the environment section in the compose yml is easiest imo. It doesnt have enough env variables to warrent an env file, for me at least.

[turbid loom]

I guess that explains why it crashed for me whenever i modified the composer file

[turbid loom]

eh

fun fact, that's where my shit failed

[slim current]

Show an error?

[turbid loom]

hold on

[meager moon]

The Log is More than 4 MB

Xd

I will need to find the error part only

[turbid loom]

nvm, managed to solve

[wet sundial]

#
# Fatal JavaScript out of memory: Reached heap limit
#
==== C stack trace ===============================

    deno(+0x2cb1758) [0xaaaae37d1758]
    deno(+0x2cb0654) [0xaaaae37d0654]
    deno(+0x2cb4c28) [0xaaaae37d4c28]
    deno(+0x2d05db4) [0xaaaae3825db4]
    deno(+0x2ed024c) [0xaaaae39f024c]
    deno(+0x2ece118) [0xaaaae39ee118]
    deno(+0x2ec4250) [0xaaaae39e4250]
    deno(+0x2ec20d0) [0xaaaae39e20d0]
    deno(+0x2ea6f9c) [0xaaaae39c6f9c]
    deno(+0x3316b4c) [0xaaaae3e36b4c]
    deno(+0x42daf04) [0xaaaae4dfaf04]
Player cache directory ensured at: /usr/src/app/player_cache
Initialized 4 workers
Server listening on http://[::]:8001
Listening on http://[::]:8001/

<--- Last few GCs --->

[7:0xffff38120000] 21024699 ms: Scavenge 1982.2 (2057.4) -> 1982.1 (2068.9) MB, pooled: 0.0 MB, 35.44 / 0.00 ms (average mu = 0.710, current mu = 0.247) allocation failure;
[7:0xffff38120000] 21026553 ms: Mark-Compact (reduce) 1997.6 (2072.2) -> 1995.3 (2050.4) MB, pooled: 0.0 MB, 261.35 / 0.00 ms (+ 1441.3 ms in 249 steps since start of marking, biggest step 17.1 ms, walltime since start of marking 1854 ms) (average mu = 0.

@slim current hey you know the issue ?

[meager moon]

This is the error

[nova bridge]

I dont think the cipher was found.

Remote cipher service request to resolve URL failed with status code: 404. Response: {"error":"Not Found"}

[meager moon]

Yea

But, My Yt cipher is hosted properly

[turbid loom]

add more ram to your app

since it says out of memory

[wet sundial]

its already 8gb

[turbid loom]

your vps or the ram allocated to your js project?

[wet sundial]

no just host a lavalink and the yt-cipher

[turbid loom]

You said it's already 8 gb and i asked you

8 gb what, the ram of your vps or the one yOU gave to yt-cipher?

[wet sundial]

ram of the vps

[turbid loom]

anyway, add this --v8-flags="--max-old-space-size-8192" if you want to give it 8 GB

or just change to 4096 for 4 GB

[wet sundial]

how much needed?

i did set the still its show heap limit

[turbid loom]

Odd, usually this process resolves that issue

[nova bridge]

not really sure how to help other than saying it is a configuration issue. I changed how my url and password were formatted in my app.yml and it worked.

you are welcome to use my cipher if it comes to it.

[slim current]

Make sure you are running the newerst version of yt-cipher. Older versions dont have the resolve_url endpoint and will 404

[slim current]

are you sure this is the newest version?

and how many active players do you average? Are we talking in the thousands?

Only time I have seen that happen is on an old version that had a memory creep over thousands of decodes

[meager moon]

Already... Using the latest version

[slim current]

If using docker, what is the image sha?

[eager pewter]

I have found an issue with v1.14.0 that songs start but it's not audible and to fix this I need to make the bot disconnect and play again or simply change the vc region.

I'm confused whether it's related with YT-Cipher or something else?

[slim current]

Likely not, but if you go back to the older versions does it still happen?

[eager pewter]

I'mma try with latest cipher and 1.15.0

[slim current]

Oh wait, on 1.14? my bad I thought you meant 1.15 already. Umm with 1.14 I dont think there is any change that would cause that issue, in theory maybe 1.15 could if something in url construction was overlooked, but my bot which plays about 30k yt tracks per day has been using 1.15 for a few days and no complaints of it so far on my end. Not impossible, but i have heard that issue of not starting a track for years every so often

[naive wagon]

yeah this issue has started on my end as well.

[eager pewter]

any fixes?

[naive wagon]

will have to try the new ones.

to see if it gets fixed

somehow

[eager pewter]

hmm

[eager pewter]

upgraded to latest version but still same issue

I have attached the screenshot, in both screenshots track was started but the bot isn't showing in the voice debug

then after changing vc region or making the bot disconnect and playing again worked

@naive wagon which lavalink wrapper are you using?

I think the issue is from Lavalink or maybe the library because the plugin is working fine.

[slim current]

I think the issue is more with discord, the LL client or LL itself rather than yt-cipher. I've heard of this issue long before yt-cipher tho

[eager pewter]

well i've fixed this by sending op: 4 event on player create with some retry and verification logics

[slim current]

What client lib do you use? I wonder if that issue comes from improper client library handling

[eager pewter]

It's kazagumo

I think it may occur due to high latency in the host maybe not sure but it's fixed now

[slim current]

Nice, might be worth writing down what you did or something, I see that question a bit, would be nice to have something for them to try

[eager pewter]

This is the code I'm using, it just force the voice state update event on each player create to the websocket in case the bot haven't connected with the RTC.

It stops retries once the bot has properly connected.

[timid garden]

If the internet worked like crazy, this would happen in most cases 🤣 else discord api 🧐

tbh it has nothing to do with cipher

[eager pewter]

yeah

[radiant ermine]

dump

[crude nymph]

Hello @slim current, I found a bug concerning the clearing of old players after x days. It still does not delete the old files. In the meantime, I restarted, but since adding it, there has been a slight bug. Am I the only one with this issue ?

Or is it that if the old player lasted more than x days, it gets deleted ?

My mistake, I found the issue that was causing the problem. Sorry for the inconvenience.

[wanton drift]

Unsure if anyone else randomly started getting these errors, but I updated to the newest lavasrc, youtube-source, and yt-cipher and weirdly yt-cipher is throwing 404 not found errors:

2025-10-15T23:51:42.379Z ERROR 404838 --- [Lavalink] [back-1-thread-1] c.s.d.l.t.p.LocalAudioTrackExecutor      : Error in playback of jYdaQJzcAcw

com.sedmelluq.discord.lavaplayer.tools.FriendlyException: Something broke when playing the track.
        at com.sedmelluq.discord.lavaplayer.tools.ExceptionTools.wrapUnfriendlyExceptions(ExceptionTools.java:45) ~[lavaplayer-2.2.3.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:121) ~[lavaplayer-2.2.3.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.player.DefaultAudioPlayerManager.lambda$executeTrack$2(DefaultAudioPlayerManager.java:339) ~[lavaplayer-2.2.3.jar!/:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:840) ~[na:na]
Caused by: java.io.IOException: Remote cipher service request to resolve URL failed with status code: 404. Response: {"error":"Not Found"}
        at dev.lavalink.youtube.cipher.RemoteCipherManager.validateAndGetResponseBody(RemoteCipherManager.java:178) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at dev.lavalink.youtube.cipher.RemoteCipherManager.resolveUrl(RemoteCipherManager.java:157) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at dev.lavalink.youtube.cipher.RemoteCipherManager.resolveFormatUrl(RemoteCipherManager.java:66) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.loadBestFormatWithUrl(YoutubeAudioTrack.java:225) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.processWithClient(YoutubeAudioTrack.java:151) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.process(YoutubeAudioTrack.java:97) ~[youtube-plugin-e10cb6a8bd8ff2b80da9303bb0a6c740c3ea9e2c.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:109) ~[lavaplayer-2.2.3.jar!/:na]
        ... 4 common frames omitted

@slim current I reverted back to older versions that previously worked for me and they are all throwing the same as well which is odd

[silver salmon]

i already got them.

[wanton pendant]

Me too

[silver salmon]

instead of a 404 im getting 500 from youtube.

        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:642) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:1583) ~[na:na]
Caused by: java.io.IOException: Invalid status code for player api response: 500
        at com.sedmelluq.discord.lavaplayer.tools.io.HttpClientTools.assertSuccessWithContent(HttpClientTools.java:168) ~[lavaplayer-2.2.4.jar!/:na]
        at dev.lavalink.youtube.clients.skeleton.NonMusicClient.loadJsonResponse(NonMusicClient.java:66) ~[youtube-plugin-1.15.0.jar!/:na]
        at dev.lavalink.youtube.clients.skeleton.NonMusicClient.loadTrackInfoFromInnertube(NonMusicClient.java:139) ~[youtube-plugin-1.15.0.jar!/:na]
        at dev.lavalink.youtube.clients.skeleton.StreamingNonMusicClient.loadFormats(StreamingNonMusicClient.java:35) ~[youtube-plugin-1.15.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.loadBestFormatWithUrl(YoutubeAudioTrack.java:214) ~[youtube-plugin-1.15.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.processWithClient(YoutubeAudioTrack.java:151) ~[youtube-plugin-1.15.0.jar!/:na]
        at dev.lavalink.youtube.track.YoutubeAudioTrack.process(YoutubeAudioTrack.java:97) ~[youtube-plugin-1.15.0.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.DelegatedAudioTrack.processDelegate(DelegatedAudioTrack.java:25) ~[lavaplayer-2.2.4.jar!/:na]
        at com.github.topi314.lavasrc.mirror.MirroringAudioTrack.process(MirroringAudioTrack.java:56) ~[lavasrc-plugin-5.11.0.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:109) ~[lavaplayer-2.2.4.jar!/:na]
        ... 4 common frames omitted```

com.sedmelluq.discord.lavaplayer.tools.FriendlyException: Something broke when playing the track.
        at com.sedmelluq.discord.lavaplayer.tools.ExceptionTools.wrapUnfriendlyExceptions(ExceptionTools.java:45) ~[lavaplayer-2.2.4.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.track.playback.LocalAudioTrackExecutor.execute(LocalAudioTrackExecutor.java:121) ~[lavaplayer-2.2.4.jar!/:na]
        at com.sedmelluq.discord.lavaplayer.player.DefaultAudioPlayerManager.lambda$executeTrack$2(DefaultAudioPlayerManager.java:339) ~[lavaplayer-2.2.4.jar!/:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:642) ~[na:na]
        at java.base/java.lang.Thread.run(Thread.java:1583) ~[na:na]```

2 parts.

[stone terrace]

Youtube itself is having issues over 300k reports about it

[wanton drift]

same here that is also showing up now

[wanton drift]

ah prob on youtube end

[stone terrace]

its working again for me

[wanton drift]

lol yea that was quick

[spiral ocean]

https://www.fox5atlanta.com/news/is-youtube-down-hundreds-thousands-report-issues-online-video-sharing-platform

wasnt quick

lmao

was down for the last hour

still is

[wanton drift]

i meant quick as in from the time i reported the issue to it 'working' again lul

[spiral ocean]

shit still broke for me

someone getting fired fr

[wanton drift]

rip call susan

[silver salmon]

call google

[crude nymph]

its work for me ^^

[river siren]

Hey y'all, I'm using yt-cipher for my music bot and still getting the "please sign in" when I try to play a song. Common issues says I have to enable OAuth. How do I go about this?

[velvet merlin]

https://github.com/lavalink-devs/youtube-source?tab=readme-ov-file#using-oauth-tokens

[river siren]

so even if I'm using yt cipher I still have to add the requestToken and skipInitialization lines to application.yml?

okay thanks

[velvet merlin]

refreshTokens*

skipInitialization: true
If you alr have a refreshToken then

[slim gate]

First run:

• do not specify refreshToken
• do not specify skipInitialization
  once you have completed the oauth flow and linked an account, youtube-source will log your account refresh token in your terminal output. You can specify this in your application.yml to avoid doing the oauth flow again

skipInitialization is optional with a refresh token. If you supply a valid token, it won't prompt the oauth flow at all. If it's invalid or not supplied, initialization will begin unless skipInitialization is set to true.

[river siren]

Managed to make it work. Thank you so much

[sacred harbor]

This is probably a stupid question, but i don't have the time to keep up with the changes and couldn't find it in search: What is 'encrypted_signature' and am I supposed to provide it myself or something? I updated cipher and yt-plugin like two days ago and problems started popping up.

Lavalink 4.0.8
youtube-plugin 1.15.0
yt-cipher should be the latest (idk an effective way to update shit in docker other than git pull && docker compose up -d)

getting friendly exception in lavalink, which boils down to:

2025-10-17T19:08:24.307+02:00  INFO 1025289 --- [Lavalink] [  XNIO-1 task-3] l.server.io.RequestLoggingFilter         : PATCH /v4/sessions/c94mruxkienr9gc6/players/477930709030404117?noReplace=false, client=127.0.0.1, payload={"track":{"encoded":"QAAA7gMASVbDhkIgLSBSw5NBIChMSVZFKSB8IEljZWxhbmQg7aC87beu7aC87be4IHwgR3JhbmQgRmluYWwgfCBFdXJvdmlzaW9uIDIwMjUAF0V1cm92aXNpb24gU29uZyBDb250ZXN0AAAAAAAC0qgAC2M3M0x4MVFVWlpBAAEAK2h0dHBzOi8vd3d3LnlvdXR1YmUuY29tL3dhdGNoP3Y9YzczTHgxUVVaWkEBADBodHRwczovL2kueXRpbWcuY29tL3ZpL2M3M0x4MVFVWlpBL21xZGVmYXVsdC5qcGcAAAd5b3V0dWJlAAAAAAAAAAA="}}
2025-10-17T19:08:24.414+02:00 ERROR 1025289 --- [Lavalink] [back-1-thread-1] c.s.d.l.t.p.LocalAudioTrackExecutor      : Error in playback of c73Lx1QUZZA
.
.
.
Caused by: java.io.IOException: Remote cipher service request to resolve URL failed with status code: 400. Response: {"error":"Invalid request body: 'encrypted_signature' is missing or invalid"}
        at dev.lavalink.youtube.cipher.RemoteCipherManager.validateAndGetResponseBody(RemoteCipherManager.java:183) ~[youtube-plugin-1.15.0.jar!/:na]
.
.
.

config file: https://pastebin.com/FMHRNXwX

[slim current]

docker-compose pull should grab the newest image

That error I think is from a slightly older version of yt-cipher when using 1.15.0

[sacred harbor]

That seems to fix this issue, thank you!
Now i only sometimes (rarely?) get java.net.SocketTimeoutException: Read timed out, but I suppose that's not yt-cipher related anymore.

[slim current]

Hmm, do you know how often? You can try adding this to the LL config

lavalink:
  server:
    timeouts:
      connectTimeoutMs: 10000
      connectionRequestTimeoutMs: 10000
      socketTimeoutMs: 10000

I think I have seen some other people with ipv6 on hetzner getting those

[wanton drift]

java.io.IOException: UT002002: Channel is closed
        at io.undertow.websockets.core.WebSocketChannel.send(WebSocketChannel.java:354) ~[undertow-core-2.3.13.Final.jar!/:2.3.13.Final]
        at io.undertow.websockets.core.WebSockets.sendInternal(WebSockets.java:906) ~[undertow-core-2.3.13.Final.jar!/:2.3.13.Final]
        at io.undertow.websockets.core.WebSockets.sendInternal(WebSockets.java:900) ~[undertow-core-2.3.13.Final.jar!/:2.3.13.Final]
        at io.undertow.websockets.core.WebSockets.sendText(WebSockets.java:63) ~[undertow-core-2.3.13.Final.jar!/:2.3.13.Final]
        at io.undertow.websockets.core.WebSockets.sendText(WebSockets.java:50) ~[undertow-core-2.3.13.Final.jar!/:2.3.13.Final]
        at lavalink.server.io.SocketContext.send(SocketContext.kt:165) ~[classes!/:4.1.1]
        at lavalink.server.io.SocketContext.sendMessage(SocketContext.kt:149) ~[classes!/:4.1.1]
        at lavalink.server.io.SocketContext$WsEventHandler.gatewayClosed(SocketContext.kt:228) ~[classes!/:4.1.1]
        at moe.kyokobot.koe.internal.EventDispatcher.gatewayClosed(EventDispatcher.java:47) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.gateway.AbstractMediaGatewayConnection.onClose(AbstractMediaGatewayConnection.java:141) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.gateway.MediaGatewayV8Connection.onClose(MediaGatewayV8Connection.java:168) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.gateway.AbstractMediaGatewayConnection.close(AbstractMediaGatewayConnection.java:100) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.internal.MediaConnectionImpl.disconnect(MediaConnectionImpl.java:70) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.internal.MediaConnectionImpl.close(MediaConnectionImpl.java:250) ~[core-2.1.1.jar!/:na]
        at moe.kyokobot.koe.internal.KoeClientImpl.destroyConnection(KoeClientImpl.java:44) ~[core-2.1.1.jar!/:na]
        at lavalink.server.io.SocketContext.destroyPlayer(SocketContext.kt:137) ~[classes!/:4.1.1]
        at lavalink.server.io.SocketContext.shutdown$Lavalink_Server(SocketContext.kt:201) ~[classes!/:4.1.1]
        at lavalink.server.io.SocketServer.afterConnectionClosed(SocketServer.kt:168) ~[classes!/:4.1.1]
        at org.springframework.web.socket.handler.WebSocketHandlerDecorator.afterConnectionClosed(WebSocketHandlerDecorator.java:85) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at org.springframework.web.socket.handler.LoggingWebSocketHandlerDecorator.afterConnectionClosed(LoggingWebSocketHandlerDecorator.java:72) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at org.springframework.web.socket.handler.ExceptionWebSocketHandlerDecorator.afterConnectionClosed(ExceptionWebSocketHandlerDecorator.java:78) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at org.springframework.web.socket.adapter.standard.StandardWebSocketHandlerAdapter.onClose(StandardWebSocketHandlerAdapter.java:144) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at io.undertow.websockets.jsr.UndertowSession.closeInternal(UndertowSession.java:236) ~[undertow-websockets-jsr-2.3.13.Final.jar!/:2.3.13.Final]
        at io.undertow.websockets.jsr.UndertowSession.close(UndertowSession.java:195) ~[undertow-websockets-jsr-2.3.13.Final.jar!/:2.3.13.Final]
        at org.springframework.web.socket.adapter.standard.StandardWebSocketSession.closeInternal(StandardWebSocketSession.java:237) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at org.springframework.web.socket.adapter.AbstractWebSocketSession.close(AbstractWebSocketSession.java:144) ~[spring-websocket-6.1.9.jar!/:6.1.9]
        at lavalink.server.io.SocketContext.closeWebSocket(SocketContext.kt:208) ~[classes!/:4.1.1]
        at lavalink.server.io.SocketContext.closeWebSocket(SocketContext.kt:212) ~[classes!/:4.1.1]
        at lavalink.server.io.ShutdownHandler.run(ShutdownHandler.kt:13) ~[classes!/:4.1.1]

new error I've come across with latest version, but seems to have not happened since around 2023

#1130136181082947664 message I've read topis explanation for it and am gonna deep dive a bit more and see wat could be causing it on my end, however if anyone has any insight it would be great as well 😄

[spiral ocean]

this has nothing to do with youtube

the error is the discord voice server ending

[wanton drift]

hmmm

ill make a new help post for it

[slim current]

new player script incoming that will require a people to pull a new version of yt-cipher soon. Ill work on it when im home in a like 5 or 6 hrs

[naive wagon]

a small confirmation .. has it fixed that issue for you completely?

[eager pewter]

yes

[naive wagon]

i might try implementing this thnx

[proper shore]

Do you plan to implement automatic updates?
probably, many people would want that

[slim current]

Auto updates? Not super feasible without restarting the process and pulling the update. That's the unfortunate reality of self hosting, luckily it's super quick and the service spins up in just a few seconds so a simple docker-compose pull and down/up is all you need. You could write a basic bash script to do that command and if it pulls a new image then reboot, put it on a 1hr timer or smth, that's a workaround for auto updates I guess

"gptcoder" just ask an AI to do the bash script should be simple enough for them to handle

[scenic fox]

it could be worse

you could have to restart your lavalink

[honest tartan]

iam getting this now

[proper shore]

I was assuming a restart would be necessary.
I'll give it a try later, Thank you!

[slim current]

ill post aometning when its done

[slim current]

what version of yt source and do you have a remote cipher service setup? It doesnt look like it at first glance

[honest tartan]

i use 1.13.5

yts.version: 1.13.5

[slim current]

go to 1.15.0 and check out https://github.com/lavalink-devs/youtube-source?tab=readme-ov-file#using-a-remote-cipher-server

[slim current]

Newest build of YT-cipher fixes this bad script 638ec5c6

[scenic fox]

yt-dlp ejs 0.1.0 release 👁️

[slim current]

yeah next thing is to see if we can use their recent changes for getting it release ready to just pull in the js as a dep rather than git cloning it 😂.

[fathom acorn]

now we need pull code newest and docker down and up sir?

[slim current]

if you are using docker just pull the latest image

[naive wagon]

@slim current i'm using deno so please update the readme for the newest build the git checkout part

[slim current]

I wont update the readme as it will just change when they merge into their master branch (when they do I will update the readme)

What you can do is look at the steps taken in the docker file and follow those or wait for it to merge (i would imagine soon)

Basically just download ejs from this fork instead
git clone https://github.com/seproDev/ejs.git
then checkout this branch in the ejs
cd ejs && git checkout fix-sig
Then just patch it as normal and run it

[wet sundial]

suddenly i getting this @slim current

[slim current]

Did you do anything to cause this? Update yt-source?

Newer yt-souce 1.15 I think uses a new endpoint that you may need to update yt-cipher to get if you use an old yt cipher

[wet sundial]

i use latets comming of yt-source

[wet sundial]

i just host new

[slim current]

I dont understand. Did you do anything when this started? Or like "I woke up and it broke randomly in the night" kind of thing?

My first guess with such low amount of details is, you updated to newest yt-source plugin but self host an old version of yt-cipher, but I cant know without more info

[wet sundial]

i just purchase the new vps and go to your github clone it host it

[naive wagon]

oh gotcha thnx

[wet sundial]

no old version

[slim current]

how are you running yt-cipher, also check your LL config maybe you are pointing to the to the wrong address or something, resolveUrl this is the newer endpoint not in old yt-cipher versions. Double check the remoteCipher url in your config and also make sure you pulled the newest docker image / run the latest commit on master from the gh repo (if just using deno)

[wet sundial]

when i using ipv6 the LL config look like this url: "http://[2a01:xxx:xxxx:xxxx::1]:8001/" ?

[slim current]

3 fixes, use the newest yt-cipher, or downgrade your yt-source, or use the public one I guess

[wet sundial]

i host in deno its fix now

👍

[eager pewter]

do I need to increase or decrease more?

[crude nymph]

Yes or improve your connection

[slim current]

I have never seen a request take more than maybe 2 seconds, and thats on a potato that is doing script prepossessing, so literal worst case, normally your looking at 2ms for preprocessed and maybe 500ms for the first request for a script. Increasing or decreasing won't help you here. Correct me if I am wrong but here is what I remember of you hitting this issue for a while

• IPV6
• Hosting on the same machine LL is on
• Tried upping timeout
• Doesn't happen all the time

Does this ever happen if you use another service like the public api?
Did you ever run the curl commands (I think I sent you a while back) to see if any of those timeout?
Did you ever try doing this on a machine from another host?

If I had to guess, its most likely some ipv6 networking setup thing on the machine. Maybe a version of yt-source with some extra logging is needed

[eager pewter]

hmm, this doesn't always happens, usually in 1 out of 100 requests

[spiral ocean]

prob buns network

but at the same time could just yt lol

ever have a yt page be slow etc

just normal network shit

[sand harness]

@fierce tusk \

[turbid loom]

yt cipher is a separate thing from spotify tokener

check the pinned text

[onyx bronze]

thats a different project btw from this, contact me if there are any problems or open issues

[scenic fox]

this thread is about the "official" cipher service

so if you use anything else, not in here

doesnt matter if its mentioned or not

you can just use the official tho

[north ferry]

its fine, it got fixed.

[nova bridge]

is there anyway to make yt-cipher use less ram?

green line is yt-cipher, blue is lavalink

[slim current]

You could run it on less threads, by default it spawns 1 thread for each thread on the host, you can set it with the MAX_THREADS env var I think. 1 is fine unless you are serving like many hundreds if not 1k+ players. Less workers will make an impact on mem usage

Unfortunately though each player script is like 20-40mb of ram use to emulate iirc. I can look more into some more efficent caching, but it may require some ejs changes

[spiral ocean]

1gb isnt bad lol

yk just let it use 1 million

[nova bridge]

its taxig me 😭

i reduced it

only uses like 250 now

[slim current]

I played around a little with adding tree shaking to the EJS preprocessing step. It added ~7% preprocess time for a ~13% mem savings so not too bad. But those solver functions really touch almost everything in the script if it only was able to drop 13% of the script file. Even though it worked, it also could not account for dynamic accesses, so in theory if they started doing dynamic accesses in their obfuscation it could break. So it's probably best to just leave it out. RAM usage is not too bad, although in this economy

[scenic fox]

how long is stuff cached for anyway?

dunno how big the cache is anyway

[slim current]

Until either process restart, or max cache size is reached. Its just a LRU cache with no TTL

Default is 150 scripts, which is really high. Could def lower it. I doubt anyone reaches 150 though. Even my public one which touches many regional scripts maybe hits 60-70 before the deno garbage men come around and clear it out

[spiral ocean]

prob doesnt reach more then 3 at once

[scenic fox]

I'd prob add a ttl

[spiral ocean]

once a new script comes out the old one is done forever and ever

[scenic fox]

but then, if deno gcs the map its prob fine

[slim current]

For sure, haven't added a TTL yet because 1. its not that much mem, and 2 deno GC seems to handle it when you start using a lot of mem

[scenic fox]

meh, idk kinda useless to cache a script which as no usages for 24h

[slim current]

For sure, haven't had a need to add that complexity yet though. The cached scripts themselves I think clear after 7 or 30 days from FS

Script is cached to disk for some amount of days, the AST processed scripts are in mem in the LRU

[scenic fox]

would be interesting to see whats the biggest memory hog in this case

[slim current]

Looking at my pub node it peaked yesterday at 2 worker threads, 110 cached scripts, 2.5GB mem usage then after GC it went down to 30 scripts and 900Mb

[scenic fox]

yeah you should definitely consider running some more aggressive ttl then

i mean its fine if you have the ram, but not everyone does
but then
they also can just use your public node lol

[slim current]

Right now if you set a lower cache or worker thread count it can run really lean

esp if you just serve 1 region, so you are always serving en_GB or something

each worker thread has its own instance of the AST tools used by ejs iirc, which is also a bit of a hog.

I will probably just drop the defaults to like 4 worker thread max, and cache at like 15 by default or smth

I was trying to figure out the est RPS per active player and its so damn low even 1 worker thread is probably fine unless you have a huge bot

1 player loading a vid every 3 mins takes 180 players to reach even 1 RPS, and when even my slow ass arm cores serve an avg response time of 1-2ms cpu time. I think it's fine for anyone short of like the biggest ones

[slim current]

Thank you for this in depth question seeking help. I can let you know, this error code has nothing to do with yt-cipher. it is common for many reasons, ranging from private video, removed video, region lock, age lock, account flagged, ip flagged, etc. This error comes before YT-cipher does its main cipher work anyways iirc

[still spoke]

Hello, is there a bug with the plugin or something else? I restarted everything 5 minutes ago, relaunched my bot, and when I try to play music, I get a bunch of stuff in Lavalink, for example, "suspicious exception".

[eager pewter]

try changing your refresh token

[still spoke]

I just changed the YouTube token, I restarted and tried again, and it's still showing me all this.

[eager pewter]

Are you on the latest version of YouTube plugin?

[still spoke]

YouTube plugins 1.16.0

It was working perfectly, I stopped to update the discord.js dependency, and since then it's been doing this: I have to search for a title and launch it, or launch the playlist directly 😅

[fathom acorn]

I'm also having a problem using individual links like https://www.youtube.com/watch?v=5RpDkUT4IHM it's not found, I use link playlist or play use title it works

[eager pewter]

Lmao, same issue

[still spoke]

I know commands, playlists, links I had saved, etc. Everything was working before. I updated dependencies, like discord.js for example. I'm currently checking what I updated and if I revert them to see if it works again.

[eager pewter]

Discord.js is not related with this

[still spoke]

You know what?

[still spoke]

Everything works fine before I update my bot's dependency

I also use Kikkia's link for its decoder; maybe that's the problem on her end, I don't know.

[eager pewter]

I'm using self hosted cipher but same issue.

[still spoke]

I also use Cypher from Kikkia, maybe that's the problem I'm having right now, I'm just trying to understand in general lol

[fathom acorn]

This problem I just encountered today is probably not due to the udpate package because I don't have any new updates

[still spoke]

Okay, I didn't pay attention at all. It had been a month since I restarted it, and everything was still working this morning without any errors. I think it's the Cypher side I'm using; maybe it's the creator's audience. I'm only seeing that right now. Or maybe they changed something on the Spotify or YouTube side, but I doubt it. As for Lavalink, I just restarted and refreshed the YouTube and Spotify tokens.

[scenic fox]

Instead of spamming useless messages you all you should read the POST GUIDELINES and send the things it requests

"Doesn't work* isn't useful for anyone to look into this to debug

[slim current]

These issues have 0 to do with yt-cipher, I know that was the last thing to change, but these suspicious track is not available, you must log in, etc exceptions come before yt-cipher is ever utilized to decipher anything. They come during track loading in innertube
at dev.lavalink.youtube.clients.skeleton.NonMusicClient.loadTrackInfoFromInnertube(NonMusicClient.java:148) ~[youtube-plugin-1.16.0.jar!/:na]

These errors come from primarily flagged IPs, or flagged accounts, etc. An account does not have to be banned to be flagged. From my observations running many many tracks over literally thousands of accounts. These errors are not predictable, but I notice them happen more to accounts that play a lot and mostly from a datacenter IP. If I run the node from a residential IP or its an account thats not constantly streaming they tend to go away. Thats my observations at least. Its a mix of heuristics that YT uses to flag you as "maybe a bot, but dont fully ban them yet" or something along those lines

[still spoke]

Thanks for your reply. I'll try switching accounts then. I went to the YouTube account I use for Lavalink, for example, and there's no message or anything to indicate a ban or anything, but I'll try switching to see if that helps.

[slim current]

It's not necesisarilly just the account, from my testing it seems like they take a ton of different variables into account. The ones I have played a lot with are IP and account. For example I saw 1 account that played fine from my home, but if I used wiregaurd to have the traffic go through the VPS I run one of my LL nodes on, it would not load videos in the browser.

[still spoke]

@slim current I just tested it; I changed my YouTube token to an account that's never been used. I also changed my Spotify developer account, but it's still the same after restarting and testing. Do you think it could be an IP address or something like that?

[scenic fox]

Yt has nothing to do with Spotify

Try a different server

Then you will know

He already said that it can be the IP being banned above

[still spoke]

The administrator of my web hosting provider tested to see if my IP address was banned. Apparently, there were no blocks, nothing strange. I'll change it anyway and see.

😘

[spiral ocean]

They don’t block individual IPs in the sense of u can’t access the site it’s not like a 429 they block ASNs basically the hosting provider linked to the IPs and just just spam u with sign in errors etc

[still spoke]

He did that as a test

[spiral ocean]

Dawg that test doesn’t mean nun

It just means u can hit the site

It doesn’t mean YouTube hasn’t flagged the network

If ur getting sign in issues while only hosting on a certain network it’s cause ur flagged

YouTube isn’t dumb they can see if a IP is Residential or not

[still spoke]

Do you think if I gave you a local example it would do the same thing? Sorry, I don't know much about this.

[spiral ocean]

Prob not ur local IP prob will work fine