#✨│ai-help
1 messages · Page 312 of 1
Hope you like it just as much as deiteris ^^
Depends on what it is tbh, applio I'd say is easier to run on cloud like kaggle, but a real-time voice changer is much easier to run locally
What's it do? I don't think I've heard of it
Pretty sure if it's not specifically made to run on Nvidia only should run fine on amd as well
Is it like Sora or kling?
I'm not sure, I'd ask around tho if anybody here has used it
sry to bother you again, but do onnx models not work with tg fork? .py works just fine, but onnx never
and it always resets the sample rate to 44100 which is a little annoying
I only use pth and index files so idk
oh okay, and does it reset the sample rate for you too every time?
it does yea
dunno why
apart from that it works just fine, so ty for recommending it. I'll just stick to pth models for the time being, there are currently open issues on the github page regarding onnx models not working, so maybe it will be fixed soon
I'm thinking of doing a double GPU Setup and using my secondary GPU for the AI voice to offload the huge load of both processing a game and processing my voice from my main GPU. Is that a bad idea or a good idea? btw I have a RX580 SE as my secondary GPU (I have the amd64-dml version).
im good aswell
i did changwe my user and displaty so i do hope u rmb me 😓
I
uh
don't lol
@hallow thistle
Why doesn't the AI model I'm using record my entire sentences? Pls help
gng twin help me
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
Tell your:
- Full GPU Name: (e.g.,
NVIDIA RTX 4060 8gb vram desktop) - Operating System: (e.g.,
Windows 11) - Detailed Description: What were you trying to do and what went wrong?
- Tutorial Used: Link to the guide you were following.
- Screenshot: A picture of the full error message.
To maintain a legal, safe & ethical community, we will NOT provide help for:
- ANY illegal activities.
- NSFW/Porn.
Requests for these topics may be ignored, not helped and result in moderation action.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
he my go to helper twin
can we add audio samples in model maker role submit application
how there is no option to add audio
where can i see that channel
i got an error when i did that
There's Applio RVC as a Google Colab notebok.

nvm its got solved twin
Can you share the link plz? 
yoo twin where do model makers upvote
You asked me like if I know how to upload a voice model to #1175430844685484042, and this is like the day third you keep pinging me just for this. 
bro we have problems😭🙏
Tell your:
- Full GPU Name: Intel(R) UHD Graphics
- Operating System: Windows 11
- Detailed Description: When installing the voice models (at 3:53), I needed to put the index file in the logs folder but there wasn't one. It was only the json and pth file. Where do I put the json file?
- Tutorial Used: https://youtu.be/_V15jq9zSJw

There's no reason to expect me to answer you about everything. While I only know about certain parts of AI programs, it doesn't mean I would know about literally everything on Earth. Some other Helper staff members might rarely go in here for some reason. Generally, if you want more helpful information, use @ helper role or ask a fellow "model maker" member if they might know how to upload a voice model in this server. Sorr, if this sounds mad to you, but I have something else to do.
💔🥀ok bruh
sry for disturbing you
i was the reason you got hurt
wtf i just got a role
what is it
Please @ me because I've got everything on mute

Sorry I meant when responding to my question 😅
hello, how do you do split lead vocals from backing vocals? im trying to make an ai cover of a song
im on vivlos169, followed absolutely zero tutorials, windows 11 5060ti, what did i do, it worked fine before
vivlos169? what's thatt
Vivlos or Vonovox? And why are you trying to run a program inside a temporary WinRAR directory?
UVR5 is a program used to separate audio stems not just vocals/instruments. https://docs.aihub.gg/rvc/resources/dataset-isolation/
Last update: November 16, 2025
is there no uvr5 for kaggle
i meant vononox mb
oh i think i figurted out the problem
i started start.bat in winrar and didnt extract it
Try extract Vonovox to a directory like "D:\Vonovox".
it worked, thanks
Is there a tutorial for the voice charger? It’s changed a lot from YouTube videos that show it off from like three years back but that’s about it
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
how can I disable the audio separation in rvcAIcover colab??
I already have separated stems
There's Applio RVC.
the quality of the output is so sad 🙂
More like low quality.
Which real-time software is the best? Mine doesn't sound the same as when I'm using it.
@hardy crow The AI HUB Docs contain both Wokada Deiteris and Tg-develop fork, they ar 2 different forks, just that Tg-develop still gets some QOL updates unlike deiteris at the moment so that's why Tg-develop is more suggested, tho Deiteris is still kept for now
You're using Weights' Replay? maybe ask in: discord.gg/weights
@oak plank did Weights lost vanity?
Are you looking for an RVC (Retrieval-based-Voice-Conversion) realtime voice changer program? Could you please tell your pc gpu and OS?
It should work now, no idea how it got lost lol
can anybody help
hi are u able to helps
What are you looking for?
to download the voice changer
but i dont seem to know what to download
What is your PC GPU?
idk how to check in macbook
its a m3 if that helps
brb i have to afk for awhile my bad
Try Tg Develop's W-Okada fork. The version b2364 is the only known version to work with Apple Silicon's Mac, though the program would only run on CPU. https://github.com/tg-develop/voice-changer/releases/download/b2364/voice-changer-macos-arm64-cpu.tar.gz
!help
Unlock the world of LunaBotPrime, absolutely free! Dive into the realm of premium music quality, enjoy it to the fullest. LunaBotPrime - where the music never stops, and the magic continues!
You can invite LunaBotPrime with this link
LunaBot 🌙 is the perfect music bot! Feature rich with high quality music! And Custom Playlist
You can start listening music by just joinning a voice channel and typing: /play [song name or link] (Remove brackets).
We support only Spotify, soundcloud, bandcamp and more!
To view more help on a specific command or category, run
/help <command> or /help <category>
Important Links:
Support
Premium
Invite
Command Categories:
🎶: Music
💰: Premium
⚙️: Utility
📕: Admin
Select A Page From Dropdown Menu Below
Did you download zip file? https://github.com/dr87/Vonovox/archive/refs/tags/v1.6.9.zip
i haven't downloaded anything yet

Where do I download this? There are way too many versions and links
It's in the link I sent to you earlier your message.
Alright I’m installing it rn, I run the setup first right?
Yes.
wow this ai is good thanks
how did we do this im back after a long time
i want to make ai vocals
i know the voice models but how can i use one
i tried the silicon thing but it said somthing abt corrupted file
ok i try again
Last update: November 22, 2025
it says "“Python.framework” is damaged and can’t be opened. You should move it to the Trash."
do i cancel first
do i copy the whole thing
Use single reply instead of having to reply the same message with multiple separate ones.
ok sorry
I'm not sure what to tell about Apple Mac since I never use this one. By the way, open a Terminal application, type xattr -dr com.apple.quarantine "a directory of your MMVCServerSIO folder", what I mean about "a directory of your MMVCServerSIO folder" is basically a path to your MMVCServerSIO (voice changer) folder. After that, try run MMVCServerSIO app again.
i accidentally blocked microphone access and since how the webview works in wokoda idk how to re enable it anyone know how to help
it says process completed but when i open the same message pops up
Review your step or try screenshot your screen.
so first i try to open MMVCServerSIO then it asys pyhton framework is damaged. I see a terminal so i just pasted that and where to find it. then it says process completed and when i retry opening the application it doesnt work and says the same thing
You sure the path to MMVCServerSIO was the right one?
yes it says
Saving session...
...copying shared history...
...saving history...truncating history files...
...completed.
[Process completed]
hey, any suggestion for setting chunk and extra for 16gigs ram and rtx3050? currently sometimes its smooth and ended with robotic voice on 96000 Chunk (2sec) and 384000 (8sec)
I THINK IT WORKED
Hello
can u help me with the virtual cable thing
Hy
i just finished extracting the zip file, and when i try to open the start_http file it doesnt install, like the black tab just pops up and just doesnt do anything, can anyone help?
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
Both Tg Develop's W-Okada fork and Vonovox can be competitive at audio quality. While I still have never got to see any benchmark about two programs on the same system, some people say Vonovox can give audio quality than other voice changer versions.
could u help me?
If you were watching a youtube video tutorial for a Realtime Voice Changer for Roleplay, it's an outdated one since start_http.bat is from the original wokada (which gets only Beatrice and UI updates, but not RVC, Retrieval-based-Voice-Conversion, which is the most popular and used type of models)
Please check #1402790586028789830 and elaborate more your request, there are multiple options depending on the hardware
For virtual cable on Apple Mac, there are Blackhole and VB-Cable.
What is your PC GPU? There's a newer voice changer version you might like to try.
So how to use this 🧐
ii need help getting the ai voice model into a voice hcnager
ryzen 7 and gtx 1650 ti
Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397
Make sure to read help guidelines before start asking, and see my question in #✨│ai-help message.
alright ty ill try it tom ill let you know if i have anymore provlems
You still got NVIDIA GeForce GTX 1660 Ti or another newer GPU?
Try Tg Develop's W-Okada fork or Vonovox instead of original W-Okada versions.
Welcome back. There's Applio RVC for doing AI cover.
-rvc
Intel Arc Axxx/Bxxx, Intel Arc integrated (within Intel Core Ultra CPU) or Intel UHD Graphics?
any intel gpu
No, but like which specific one? One being a dedicated GPU, while other two are integrated GPUs.
nevermind figured it out
change models then idk
Guys how to fix robotic & cracked voice of W-Okada? Im trying egirl voice.
Using RTX 2060 SUPER.
okayy im installing Vonovox right now, but its stil cracked and robotic :'3
im installing Vonovox 1.7 beta
What's thats
*Vonovox , sorry for typo : https://docs.aihub.gg/realtime-voice-changer/local/vonovox/
what gpu you're using?
im facing robotic sound ah
With rtx2060s
im using 3050 6gb
same as me 🙂 its crack and sounds robotic
i have dmed you. if you find any fix lemme know🙂
okayy
can someone help me make my egirl voice changer less laggy and more natural (w-okada) ?
Does anyone know a girl's voice that is very realistic and good for the German language?
Pls Help
-rt
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
I'm having some truly annoying problem in that an installed node in ComfyUI is just absolutely refusing to show up in the Node Library.
Installed and enabled as it should be via the custom node manager.
But just. Not showing up???
why is my okada so delayed did i download the wrong thing?
does okada work with a intergrated gpu
I use the Tg Develop W-Okada Fork application and everything is fine in server mode, but when in client mode my mic doesn't give input, why is that?
I felt crack and robotic from server mode so I wanted to try client mode
Really need help for this one
no idea, itt doesn't work for me either
how do i fix cutting out
Dont updated yet
@hallow thistle when you get a chance could you help me with this?
how would you be able to solve the problem without seeing error messages in the console window?
I have no clue about that comfyui node so you'd better pass the error messages to Claude, etc.
Hi, Is there a voice-over feature for movie subtitles in real time? Preferably a local one for a PC. Or an audio translator from one language to another?
I would like to transfer a few local movies on my PC to another language
Hi... Can someone help me with the first steps? Because I'm just interested in using Voice Changer in game
what should i use? Applio? Okada?
applio is for training voice models
That means, I have to use Okada? because im not training Voice Models
ye, also depends on what gpu u have
i have rtx 4070, it is okay to use Okada?
and this is so much Okada version, what should i use?
ye, you could use either vonovox or wokada tg fork
yeah, i will using Okada, but what version of this works? ...
both versions I mentioned work
You have stated your PC GPU is Intel UHD Graphics, an integrated GPU, so while any W-Okada version (b2397) with DirectML might work with this integrated Intel GPU, the performance would nowhere be the level as a dedicated GPU (like Intel Arc Bxxx). If there's another GPU in the same system, let me know, or check out online cloud services for more voice changer performance otherwise.
Download and try b2397 W-Okada. https://github.com/tg-develop/voice-changer/releases/tag/b2397
What is your PC GPU? And did you follow any tutorial or guide before?
Cuda 001 or 02?
im using tutorial in #1159513888199540817
So does this mean you still got GeForce GTX 1660 Ti? For more recent W-Okada version, check out Tg Develop's one. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397
Download both .zip.001 and .zip.002.
alright, lets do it, 1 hours left
||Ever since AICoverGen got discontinued I've been having difficulties working with the other options like Kaggle even with the guide.|| This video https://youtu.be/_V15jq9zSJw helped well and my laptop is running fine. I was just needed help with installing the models because the models don't have an index just a json and pth. Idk where to put the json.

If you mean a regular RVC (retrieval-based voice conversion), try delete your current Kaggle notebook and follow this https://docs.aihub.gg/rvc/cloud/applio-kaggle/ guide.
Last update: September 30, 2025
Sorry I didn't to be a bother
This one is W-Okada "realtime" voice changer. https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/ Both Kaggle notebooks need you to import from GitHub to your account.
Last update: September 6, 2025
how can i use tg okada and voicemod together with discord
As what I said in #🧬│ai-chat message, the voice changer might work with an integrated GPU but that's about it. For full performance, you might wanna check this https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/ Kaggle doc.
Last update: September 6, 2025
No idea about Voicemod, but if without Voicemod, the settings on W-Okada would look like this. https://cdn.discordapp.com/attachments/1159290139609137264/1456249974525526185/image.png?ex=69643439&is=6962e2b9&hm=9b77b4625ee1402355a585f6761d59ab59de4b6e87f06753a563929c5432d77b& The Voicemod might give you its own virtual line for use with other program(s) like W-Okada unless you know how to route them correctly.
<@&1159293204038955078> where is the model maker upvotes are given
just change ur mic in voicemod to the virtual cable u use
its on that but nothings going through

is this supposed to say line 1 or do i change it to my mic?
I use a diff setup to most people so forgot how to get it to work just with okada and voicemod
what is this
site settings
what site?
okada?
I'm not familiar with that version you're using so no idea what to put for that
oh
W-Okada | Robotic & Cracked Voice Issue
🥀GPU RTX 2060 Super 8GB
Intel Mac or Silicon (M2, M4) Mac?
Did you follow any tutorial or guide before?
Yt 🙂
Try this version. https://github.com/tg-develop/voice-changer/releases/download/b2364/voice-changer-macos-arm64-cpu.tar.gz Make sure to follow the guide doc as well, especially security part. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/#opening-on-mac
Last update: November 22, 2025
Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397
wheres the download to the ai voicechanger
What is your PC GPU?
nvidia
this is so confusing
yeh
Alr thank youu!
For a tutorial about voice changer on Apple Mac, unlikely.
THATS NOT FOR PC :<<<<
the same one from this server
Which GeForce RTX?
gt 1030
dude
@hallow thistle
i downloaded
voice-changer-macos-arm64-cpu
i opened it from a zip
im in MMVCServerSIO folder
my virtual audio output cable not doin anything cuh
bruh weres the download
DUDE
check
SOMEONE HELP ME OUT?
This GPU has 2 GB VRAM, so while the voice changer might work it would very struggle a lot. The minimum possible GPU being GeForce GTX 1050 Ti or any GeForce RTX of 8 GB. For more performance possible, consider for an online cloud service like Kaggle instead.
...
i used it before and it worked perfectly
just gimme the download
Calm down, please. I was busy helping other member.
sorry :<
yea cmon man
SHHHH 🙁
!howtoask
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
Tell your:
- Full GPU Name: (e.g.,
NVIDIA RTX 4060 8gb vram desktop) - Operating System: (e.g.,
Windows 11) - Detailed Description: What were you trying to do and what went wrong?
- Tutorial Used: Link to the guide you were following.
- Screenshot: A picture of the full error message.
To maintain a legal, safe & ethical community, we will NOT provide help for:
- ANY illegal activities.
- NSFW/Porn.
Requests for these topics may be ignored, not helped and result in moderation action.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
Tell your:
- Full GPU Name: (e.g.,
NVIDIA RTX 4060 8gb vram desktop) - Operating System: (e.g.,
Windows 11) - Detailed Description: What were you trying to do and what went wrong?
- Tutorial Used: Link to the guide you were following.
- Screenshot: A picture of the full error message.
To maintain a legal, safe & ethical community, we will NOT provide help for:
- ANY illegal activities.
- NSFW/Porn.
Requests for these topics may be ignored, not helped and result in moderation action.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
?
What is your step again? Running MMVCServerSIO app from the folder?
No. You can send a screenshot to here now.
finally
Even with an integrated GPU, it won't be the same performance as a dedicated GPU or sometimes worse.
what is this?
Use an archive extractor on your macOS to extract .tar.gz archive file.
That one is macOS Finder, an app that works similar to Windows Explorer.
Double click on MMVCServerSIO.
Why would you do this?
An "archive extractor" feature exists in macOS, there's no reason to use an online website for such simple task.
called?
Go inside MMVCServerSIO folder, double click on an app file of the same name to launch the program. If the system blocks you from running it, do this folllow.
thats what i did before
?
Namari?
@hallow thistle ???
<@&1159293204038955078>
can anyone help me?
im having an issue where the voice playback pluss using it in discord or calls is just extreamly laggy. i have a rx 7600xr, ryzen 5 2600.
YO GUYS CAN SOMEONE PLS HELP, so i have a 5070 i9 and its not working since i switched from my 3060? what do i do
@lofty tusk
pray to god
wait so it just doesnt work?
idk
DAWG
woof woof
🔥
For AMD Radeon RX, try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip
Last update: November 22, 2025
Follow this message #✨│ai-help message and make sure you don't ping random members other than "@ helper" just for this.
@hallow thistle any idea for this one? sorry for ping
idk how to do ts bro
i saw something about that earlier, i just installed it and it works better, any reason why the original app just completely dies?
Because the original W-Okada version has long been outdated, its DirectML variant is also very buggy for Intel/AMD GPU.
fair.

yo can somebody show me how to do this brooo
I'm off from my laptop screen, so I can do just this.
so like when im in the website theres nothing to download or ANYTHING
Make sure to read the guide doc as well, and don't always expect too much from here.
Use search feature on Discord or in #1192011222023950368 if necessary.
Alright so I got an NVIDIA RTX 5070 ti laptop gpu with 24 gigs of ram running windows 11 and I was trying to use RVC but I get no audio from my virtual audio cable output can anyone help
Anyone on here willing to give a rundown on any stable diffusion updates? Been seeing quite a lot of AI models looking exactly the same on different posts consistently nowadays, mans leaving money on the plate not playing that game.
what is the best voice changer to use
Hi Chat
I'd like to ask about a good mic + interface suggestion for the purposes of using RVC
I did some digging and discovered those:
Mic > xm8500
Interface > Behringer U-Control UCA222
I'm sorry that I couldn't post any pics.
Are these two good? the budget is kinda pretty low at the moment.
sorry again for the ping but do u have a video tutorial on how to use it? i already made it into the website that it directed me in but i couldnt hear the voice changer
is there a fix where the ai voice is like static? or do i have to adjust just settings
Chunk: around 160 ms, always make sure to check perf number at top right screen
Extra: 2.7 s
GPU: NVIDIA GeForce GTX 1650 Ti
Pitch extraction: rmvpe
Try Vonovox or Tg Develop's W-Okada (b2397). Most older W-Okada versions (like b2332 and v.1.5.3.18a) weren't compiled to work wirh GeForce RTX 50 series.
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
Every microphone (even cheaper one) will work with W-Okada realtime voice changer program, as long as your microphone can capture audio signal perfectly. The term "RVC (retrieval-based voice conversion)" refers to a different non-realtime program (like Applio RVC) that also uses RVC (retrieval-based voice conversion) voice models. The audio quality of voice changer has to happen with chunk and extra settings, your PC GPU, not always the microphone itself.

W-Okada voice changer (MMVCServerSIO), Vonovox and Applio RVC are three different programs. For W-Okada voice changer, download https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001 and https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002.
There are ComfyUI and Forge, by the way.
Vonovox or W-Okada fork? For W-Okada fork, here's your settings:
Chunk: around 130 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 3050
Pitch extraction: rmvpe
Input: microphone
Output: Line 1 (Virtual Audio Cable)
Monitor: your speakers or headphones
I see...
I'd like to use W-okada indeed, although for work purposes ( a small marketing agency ). The current mic I'm using - per my boss' request - isn't quite bad, but it's highly unstable and produces echoes and cuts out after prolonged usage.
It's a cheap mic called "Elegiant", and my work PC has a GPU that's 3060 12 GB from NVIDIA. It also has 32 GB of RAM, and an i5 3600 for the CPU.
So these are my work specs, and my use case.
I followed the Deiteris' W Okada Fork for the best chunk/EXTRA numbers, but the issue mentioned above was still present, even after using different models.
Which is what made me think that it's probably an issue with the microphone, being a cheap one and all.
Am I correct in my assumption that I need a better mic + interface?
And are these listed above good picks in this case?
and then i just download a modle in find modles? and i just upload it?
Do you mean third-gen or 13th-gen Intel Core i5 or an AMD Ryzen CPU? The Deiteris' W-Okada fork (b2332) was last updated in December 2024; there's a more recent and continuation one you might like to try.
For about microphone, see this part https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/#do-i-need-an-extremely-expensive-mic-for-good-quality.
Last update: September 6, 2025
Omg
I'm sorry
Yes it's ryzen 5 3600.
I'll read the link thoroughly.
Thank you
This is there you upload a voice model on the voice changer. https://cdn.discordapp.com/attachments/1159290139609137264/1458389948796375070/image.png?ex=6964bcfb&is=69636b7b&hm=a7a9a7eab879b191e9df65e7e53134923f35246a91fcc1bd5c82ef733ec1125d&
does anyone knows if there is a different launcher or app for this voice models that i can use where it does not stuttering like sometimes it cuts out?
!howtoask
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
Tell your:
- Full GPU Name: (e.g.,
NVIDIA RTX 4060 8gb vram desktop) - Operating System: (e.g.,
Windows 11) - Detailed Description: What were you trying to do and what went wrong?
- Tutorial Used: Link to the guide you were following.
- Screenshot: A picture of the full error message.
To maintain a legal, safe & ethical community, we will NOT provide help for:
- ANY illegal activities.
- NSFW/Porn.
Requests for these topics may be ignored, not helped and result in moderation action.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
What is your PC GPU? W-Okada voice changer or Applio RVC? And did you follow any guide or tutorial before?
mine works fine sometimes but the thing is i believe what im using as a app or launcher to make this work these ai models sometimes it sounds stuttering i hae a GPU of 3080 RTX on my laptop and the guide i did follow was on the youtube guy where he mentioned at one point hes being hacked by a syberian government XD when his anti virus pops out
W-Okada voice changer is a free and open-source program. If one of your antivirus programs detected it, it's always a false positive, unless you download the program elsewhere outside Hugging Face and GitHub. For better working voice changer version, try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397
alrighty i will try these thank you
it still doesnt sound like the model preview when i uploaded it
hello do you got some tutorials how to install or make it work i tried downloading it and extract it says the file was currupted or i did download the wrong thing?
You should download both voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002. These two files have to be in the same folder, use WinRAR or 7-Zip to open the .zip.001 one.
alright im going to try again
i did download them both and i was able to extract now is there a guide here which step is next?
Go inside MMVCServerSIO folder, double click on MMVCServerSIO.exe. https://cdn.discordapp.com/attachments/1159290139609137264/1459034831940747419/image.png?ex=69647293&is=69632113&hm=a38295b983127b993e13874d689bdf3b1dc092d47ba69abce278c10640b2d5b1& https://cdn.discordapp.com/attachments/1159290139609137264/1458387597209501777/image.png?ex=6964baca&is=6963694a&hm=18a42c8dee6a4ea0351875a543a7afab66c5efbb89c0168a4c41c874d9295786&
alright one sec
do you mind if i message you the ss of what im doing? like DM you so it won't bother someone that would read here?
cuz i cannot send ss here or image seems like
Talk a little bit in here until your name turns blue so you will be able to send an image here.
alrighty
uhmm i think after i double clicked it it installs something
then it shows this
oh it turns blue xD
i can start putting models here correct?
also why it was on browser there isn't an app for this?
Still expecting for a response?
The guide doc exists, though I doubt if you still reading it as a reference.
i read it earlier but it was so much information i was confuse
Here's your settings:
Chunk: 50 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 3080
Pitch extraction: rmvpe
Audio mode: server
Sample rate: 48000
Input: (WASAPI) microphone
Output: (WASAPI) Line 1 (Virtual Audio Cable)
Monitor: (WASAPI) your speakers or headphones
ok one sec
i cannot hear myself i dont know why am i doing something incorrectly?
Yes, very incorrect. Chunk should be 80 ms, Extra is **2.7 **s, not 0.3 s.
The setting "extra" means "extra processing time" in this W-Okada.
for the chunk it only moves from 2 - 3 cannot do 2.7
too low chunk, increase it till it stops showing red perf graph
First of all, click "stop server" and try set them again.
how about now?
The "Extra" part is still incorrect.
the 2.7 for extra there was no option to choose that when i moved it, it only goes from 2 to 3
Brute force it or refresh your voice changer interface.
This is the correct settings, by the way.
yes im trying but it wont gave me
i tried using mouse or keyboard to make it move slowly it goes only for 2 to 3 by one's
i think im using client earlier instead of server maybe that was the reason it wasn't working no?

when i am able to change that to 2.7 i wont hear this noises like a hum that was annoying when you got some background noises?
Hi
If you want to get help about something, say something.
hello i think i did make it work thank you the only problem im having is the background noises whenever im not speaking it murmursss sound
oh by the way what is this chunk for?
suddenly everything was not working and im getting this error i restarted my laptop then that one pops how to fix this T^T @hallow thistle
Hello
help when i open up a game like GTA or RDR2 it makes the voice changer sound stuttering and so laggy is there a fix for this?
process lasso what is that?
im trying to use this on this game i did try to make the chunk goes up to 300 hopefully it wont stutter now
ohh how i can do that is there tutorial in youtube for it?
alrighty i will take a look
Try increase chunk number up to 200 ms.
300 works for me it does not stutter much i might increase a little bit incase it stutters again thank you... so muchiee
i really appreciate your help
hey, i upgraded to RTX 5060TI and now i can't work with UVR5 is there is any new update or a work around to be able to use UVR models locally?
mainline was working fine for me yesterday and now it's broken. is there any fix for this?
I haven't used mainline in years
I prefer its inference on some things, so I still use it while using Applio as well
How does one use both?
Sometimes I inference on Mainline if the inference on Applio sounds a little off, thats all
same audio, same model
Ohh ok
yall
how do i get low delay with a amd gpu
i tried reinstalling 2 times but i get crazy delay
This guide might help you https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
update
for the record i do have a supported nvidia gpu
update 2: deleted the entire gradio folder and fixed everything 👍
also: When changing Chunk, Extra or Crossfade size settings, you must switch device to CPU then back to your GPU. Otherwise, performance issues can be observed. It's one of the known issues.
any help here please??
any heko?
any one that can help me out?
i use comfy ui
so any comfyui god would be great
This link no longer exists?
hey, i upgraded to RTX 5060TI and now i can't work with UVR5 is there is any new update or a work around to be able to use UVR models locally?
is there a push to talk on this app like its annoying about when you hear murmur sounds when it picks up the background noises like i only wanted to hear it whenever i am speaking?
You can set up a noise gate (minimum required volume before output is send), and limiter for very loud ones.
If you're not sure how these work, you can use the simple low pass and high pass for these as well.
To filter out noise, you do need an actual filter though, and so would need to run it either in client mode with noise suppression 2 enabled, or in server mode and then route the audio to something that has a noise filter. (like discord).
i know how to lower the input so others won't hear it but the monitoring device on this app whenever im not speaking and it pick ups background noises it mumurs like its so annoying do you recommend something like an app so i won't hear the mumur sounds unless i really did speak?
It depends on the usecase. For discord for example, you could raise the "Input Sensitivity" slider and enable krisp (to get rid of noise).
Sure, you'll hear it in the VC Client, but you won't in Discord (test using the output there).
For push to talk-- it kinda depends on the program as well, but there is probably a way to do that artificially. (not sure yet)
Discord has a push to talk option under Input Mode
Yeah, no clue if there is an all round working option; sorry. I know minecraft has a push to talk setting, but no input sensitivity slider, so you'd need to use the tg-develop's w-okada noise gate filter, or route your audio through something like OBS and then use a filter on that; but OBS has no push to talk button--
perhaps your microphone comes with software that can do these things as well though.
Someone else may be able to help you more than me; but you'd need to wait a bit more I guess
yes i will be needing something to filter out the voice i guess or the mumur sounds aside from that it was ok my laptop comes with ai noise cancelling but for some reason it was not working form e
so i had to use the mic itself alone
i tried downloading krisp hope it works but it was only for 7 days OMG i thought there was free one's for this
am i supposed to use the cable input thing? i cant get it to work
Krisp is build into discord. (see audio settings)
What are you trying to get to work? w-okada? one of the forks of w-okada or vonovox? (I am guessing something in that direction at least)
Also work with what? The audio signal needs to go from somewhere to somewhere
anyone have any idea why my mic is like static when i speak
when i have the voice changer on
Static? That could be an audio filter, or part of the model you're using.
hold on
It could also be caused by a bug to be honest
sigh
I don't know; you'd basically have to look at every piece in between the microphone and the output. Static can be caused by anything.
Apparently, Wave Link 2 may be causing static; which is resolved in version 3.0.0 (beta), in case you use that.
OBS can add ststic as a filter; there is also a static filter in tg-develop's fork of w-okada (not sure which RVC program you're using)
A bad connection on a usb cable can also cause static; some microphones have electric static by default due to bad isolation basically
(not saying that is going on, but- yeah- it's difficult to say which it would be)
if i use this i need to use Tg Develop's W Okada Fork?
for output realtime vc?
For a specific UVR5 release that can work with GeForce RTX 50 series, see this comment https://github.com/Anjok07/ultimatevocalremovergui/issues/1812#issuecomment-2831549418 on GitHub. Be aware that certain UVR models might not work in this version. For more information about UVR5 for RTX 50 series GPU, there's a public Discord server named "Audio Separation", you can search in "Discover" feature.
What is your PC GPU? And did you follow any tutorial or guide before?
See this message #✨│ai-help message or read help guidelines before start asking.
i did download the krisp app so far it was good but i only got the free trial one for 7 days as far as i remember there is a free of this even though not paid right and it works just fine?
W-Okada voice changer or Vonovox? And what is your PC GPU?
There's no "press to talk" feature in any voice changer (even Vonovox) like in Discord. There are potential workarounds, check "Noise Suppression 2" on and then increase "Input Sensitivity" up to around -64 dB. If your microphone still picking background noise up, also see if your system audio control panel has "Noise Suppression", or otherwise try a better, dedicated microphone instead of an integrated one.
hello thank you for this but what i did is downloaded a krisp it works fine does not even need push to talk the annoying murmurs sound is gone only when i speak it was going to activate but the thing is it says i only have 7 days wasn't there a free version of krisp even it wasn't the strong one that filter almost everything out when it was paid i need a free one T^T
Amd
Random YouTube tutorials that all said the same thing
Which AMD Radeon RX?
I forgot ngl, it’s pretty ass

To check your PC GPU, go to Task Mananger, go to Performance, see if there any GPU 0 or GPU 1.
There’s 0
Click on GPU 0 to reveal its full name on the right side.
It just says AMD Radeon
What now
Where? This one’s confusing
Last update: September 6, 2025
AMD Radeon(TM) Graphics is an integrated GPU often found in AMD Ryzen systems. The more recommended GPU, if Radeon, would be AMD Radeon RX 5000.
hello i am trying to install vox but it says i do not have a moduale installed called absl i tried pip installing but i got an error
is this still the same thing everyones using? i havent kept up with it
i saw someone say to use deitritis fork for 5000 series idk if thats still true
The b2335 was compiled for GeForce RTX 50 GPU as a temporary replacement for Deiteris' W-Okada (b2332), although there's a newer more recent voice changer version you might like to try.
Do you mean Vonovox? Try delete your current Vonovox folder, check if you have downloaded this https://github.com/dr87/Vonovox/archive/refs/tags/v1.6.9.zip, and extract to somewhere like "D:\Vonovox".
would you mind linking it?
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
Tg Develop's W-Okada fork (b2397) and Vonovox are known to work with RTX 50 series.
hi can you help me please
i have installed vonovox but it appears that i can't hear myself
While Vonovox doesn't have a second output or "monitor" like what W-Okada has, the only workaround is to set "output" to "your headphones/speakers" on your Vonovox to hear the program, and you may set output back to "Line 1 (Virtual Audio Cable)" if things right. https://cdn.discordapp.com/attachments/1159290139609137264/1446358776587489351/image.png?ex=6965cb14&is=69647994&hm=58366a64fe1c25b09afc86a3803ed2f9e6d9864f2e640a479b0932221c26beff&
it did have me install some new stuff but in the end it was the same result saying it didn't have the moduale
what is the best settings for this if i have a 5070ti
Check if you have double click on "setup.bat" file at first, and then "start.bat" the second and all. The another cause might be when you right click and run as administrator which can cause issue.
hi again sorry, so i figured it out and i can hear my self with the model talking, but i cant hear it in dicsord? could you help me please
also how can i download the voice changer client demo? like the actually software and not the website voice changer?
There's no W-Okada fork that runs as its own separate program, the original version (v.1.5.3.18a) does. The version v.1.5.3.18a is often avoided because it's outdated.
oh alright ill keep using the website but how about the discord thing, i can hear my self in the website with the model voice changer delay but i cant hear it in my discord preview
You might be thinking of this.
yes, when i did the lets check i dont hear the voice changer
what is the difference between the w-okada fork and vonovox? is vonovox faster since i have an nvidia graphics card
best e girl voice
Hi! I'm fairly new to the RVC world, started learning recently and I'm training my first models.
I use voiceover extracted from a game as dataset and before preprocessing I removed ~15% of the data by discarding the samples that don't contain actual words (grunts, laughter, death sounds etc.). My logic was that learning on those samples could cause those sounds to emerge in generated speech which would be undesired. Still, I'm not fully aware of the capabilities of the model so now I wonder, is that actually the case?
Is there any benefit to including samples that contain non-verbal sounds like grunts, screams, laughter etc? Or is my approach correct that it's better to discard those?
so what it a good voice changer client now?, i used to use the okada fork but maybe there is something better now
anyone know why w-okada isnt working for me i set the out put as the vb cable but no audio i have a 3060
I have a question regarding quality and delay for RVC (not w-okada). I found the guide at https://rentry.co/RVCRealtimeGuide#gpu-chart-for-sample-length-extra-delay-quality. It shows a table regarding the different versions of GPU's (xx50, xx60 etc). Is the values set here equal for 30xx,40xx,50xx etc or would the quality/delay improve the newer the model is? I'm asking because I have a 1060 at the moment and the delay is pretty high with this card if I want the outcome to be any good. Would a 2060 do the same as a 3060? No other applications using the GPU will be running simultaneously
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
Should I try Vonovox if Okada is simply not working/ doing anything?
I start it but dont hear anything at all even with my virtual output cables
Is there a tutorial for realtime voice in mac
Hey Im using latest Vonovox and im missing the two options below block size in Voice Settings
Were they removed?
is MMVCServerSIO outdated now?
i was following a tutorial
but couldnt get the start bat to run
hey guys so i uploaded an ai carti voice model
and its just stuttering
and not sounding right
how do i fix this
also it crashes dione
im trying to make stuff like twxn
im having the same issue, my gpu is the nvidia rtx 2060 super and im using the w-okada. plus sometimes the voicechanger doesnt even pick up my voice, though my mic and headphones are correctly connected. any help? (i didnt see a responce from the other person)
deiteris doesnt support cuda for your gpu, you should try vonovox instead https://docs.aihub.gg/realtime-voice-changer/local/vonovox/
Last update: November 21, 2025

If the W-Okada version is v.1.5.3.18a, then yes.
In Vonovox v.1.6.9, these three settings are all intact even when my laptop has no GPU.
what are people using now
whats the new W-Okada versions
ive seen people still using that version
but idk
Applio Silicon (M2, M4) Mac or Intel Mac?
What is your PC GPU?
Hey! So, I'd like to try out W-Okada on my RTX 5080, but which fork should I use?
b2397 W-Okada fork or Vonovox.
Thanks!
This is likely an integrated Intel GPU. How do you know if an integrated GPU can work with W-Okada voice changer and Applio RVC?
Unlikely.
Follow this message. #✨│ai-help message
what's the difference between vonovox and wokada?
https://github.com/w-okada/voice-changer/blob/v.2/docs_i18n/README_en.md I'm getting w-okada from here
can it do the same thing as vonovox?
This is the mainline W-Okada. In my screenshot, it's a different W-Okada version.
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
mainline w-okada still good tho right?
getting it from here atm https://huggingface.co/wok000/vcclient000/tree/main
or what do u recommend?
The version v.1.5.3.18a was released in 2023, so likely outdated.
vonovox or wokada? and which fork
What is your PC GPU?
Vonovox will work, W-Okada fork (like Tg Develop's W-Okada) will also work as well. The version v.1.5.3.18a should always avoided for most scenarios.
nivida rtx 3070
Inconclusive. For easier and more familar settings, there's W-Okada fork. Vonovox can give better audio quality than the former, but with less familar settings.
There's #1175430844685484042. RVC voice models from Weights.com will also work.
I just hit download on weights.com and I should be able to use those yeah?
with vonovox
straight away
I have to type faster to follow up your message. 
sorry mb haha
Make sure to rename the files so they're not all called "model"
okay sounds good
https://huggingface.co/dr87/vonovox/tree/main vonovox 1.7 beta yeah?
Vonovox v.1.6.9 is the most stable version, while 1.7 is beta/pre-release.
do u recommend 1.7 or 1.6.9 atm
Do I? 
Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
Last update: November 22, 2025
yeah do u recommend 1.7 or 1.6.9
Generally, I do not have any absolute recommendation for others, but for most stable working version I'd prefer 1.6.9.
out of all the assest which one do i get they are all amd??
Because so many people here have been confusing "AMD64" for an AMD Radeon GPU lately, "AMD64" is another CPU architecture name for x86-64 or x64, this CPU arch is used in most Intel and AMD CPUs but often in their different marketing CPU arch names.
If you see one that says "CUDA", it mostly refers to NVIDIA GPU. Download both voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002.
should I get vac470lite or vbcable or doesn't matter?
okay doke
Vac lite is more recommended
Over VB cable
vac470lite.
okay sounds good
vac470lite means Virtual Audio Cable lite.
Double click on "setup64.exe".
okay cheers
"setup64.exe" is for x86-64 (64-bit), "setup.exe" is 32-bit (x86) only, while "setup64a.exe" is for AArch64 CPU (like Qualcomm Snapdragon).
okay
is vonovox better or this one in ur opinion
yeah Duckus's tutorial, are there other apps I should use instead of that one?
I'm highkey confused right now, now people are saying there's different ai voice mods application thingy
which one do I use then?
I guess I'm gonna do Vonovox?
-realtime
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
I can use it on discord, no? Vonovox?
Yes, you can.
Just download and install VAC Lite:
-# https://software.muzychenko.net/freeware/vac470lite.zip
Once installed, in Vonovox, select:
Input: Your mic
Output: Virtual Audio Cable
And in Discord, select:
Input: Virtual Audio Cable
Output: Your speakers/headphones
VAC470 lite, right?
yep thanks, I just wanted to confirm it
I'm assuming the other vcclient app was an older version, hey
sorry sir, important question, I set VAC as default communication device and my laptops systems as default device, could that potentially be bad, does VAC need to be default device instead?
What do I need to make it work?
what are the equivalent settings to the "tune" and "chunk" settings on w-okada in vonovox?
when I launched start.bat, it only loaded the command promt app, not the Vonovox app
the hell
Did you run setup.bat first?
In Vonovox folder, double click on "setup.bat" first, and then double click on "start.bat" the second and all.
Vonovox is for NVIDIA GPU only. Are you sure your PC has an NVIDIA GeForce RTX GPU? Because another user didn't ask you that one.
okay it loaded after like 5 minutes
yes
I checked the site before asking here
I just got confused
how do I test it out and hear it out?
Running voice changer with an "integrated" GPU, while might work especially with UHD Graphics, is often avoided because the performance is nowhere the level of a dedicated GPU (like Intel Arc B580 or NVIDIA GeForce RTX). For more performance, an online cloud is considered.
Which GeForce RTX?
3050
Vonovox should work as well.
yeah I hoped so 😂
See my forwarded message above your query.
Yes, because b2332 W-Okada fork wasn't compiled to work with GeForce RTX 50 series, so Vonovox or Tg Develop's W-Okada fork (b2397) more preferred.
I did however I keep pressing start to test but I hear nothing
Online cloud is it like a app ?
No, it looks like this. https://www.kaggle.com/
So after I downloaded it what do I do
Why would you download Kaggle as your typical app? Follow this guide. https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/
Last update: September 6, 2025
Im confused-
Virtual Audio Cable is a program that you can download and use, with W-Okada voice changer of course, acts as a virtual line to route between programs. A voice changer (W-Okada) is a separate software that converts your voice in realtime using RVC voice models.
Its setting up
What does this mean by setting up?
Make sure to read the guide I sent to you and tell which part still confuses you.
When you say like "what should I do after I do", I'm not sure which step you're currently at.
I'm at like trying to make it wor
That doesn't refer anything.
helo
Is there something you would like to get help about?
What is your PC GPU?
rx7800 xt
The original voice changer version (like v.1.5.3.18a), especially its DirectML variant, is often avoided because it's very outdated and buggy for AMD Radeon GPU. For better working one, try Tg Develop's W-Okada fork. https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
Last update: November 22, 2025
is there any tutorial about downloading im kinda slow
Unlikely.
Better not follow any tutorial from YouTube while following the W-Okada fork at the same time.
alright
this tut uses the main outdated version of w-okada
The original mainline W-Okada voice changer is only known to work with Beatrice voice models, for some reason, meanwhile W-Okada fork (like b2397) only uses RVC voice models with Beatrice voice completely removed by default.
so which is better the new w-okada or vonovox?
Because people been asking me if Vonovox or W-Okada fork is better, to which I just said this #✨│ai-help message.
fair enough mate
quickly, are you only able to test vonovox on discord and other applications and not on the actual vonovox app itself?
While I've long been demanding for a benchmark comparison between Vonovox and W-Okada fork (b2397) on the exact system, my laptop is old and doesn't have a GPU so I not be able to test things like that.
so you haven't been able to use the apps?
can someone vc
so I can test it
on call 1
I can simply open these programs as dummy applications without actual realtime voice changing feature, so when people asking me for settings on the voice changer I would screenshot to let them know.
I don't know what I'm doing wrong for it to not work
ahhh I see, I was the same before and then I got a new laptop with a gpu so I was like "imma try this out"
In my screenshot, it's a more recent W-Okada. What you use is the original W-Okada version, which should be avoided for most scenarios.
okay what do you need help with?
Namari, in terms accessibility, which app was easier to navigate around and use?
oh okay
W Okada fork?
what
If you didn't follow a guide, you can just say you didn't follow a guide, and I'll repeat the instructions just for you.
I followed a guide from a yt guy
Basically downloaded the voice changer from him
Guides for Programs that use RVC Models in Realtime for Calls/Games
A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE
A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
So just means you didn't follow the guide I sent to you. Your PC GPU is an integrated Intel GPU, delete your current Kaggle notebook and follow this guide https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/, again. 
are you able to hop on vc to test it out? you don't need to say anything, I just want to try
I'm at work, so I don't have time or place to listen.
alright no worries sir
is there anywhere
service/provider i can hire to make custom high quality voice models for me?
Paid request and commission about voice models aren't allowed in this server anymore, but now you can make a free voice model request in #1159289738314919936.
thanks twin
question
like being so frl question
so like
the people with a high quality vc that sounds so real
do they do voice acting irl too

Most model makers here don't do their voice acting, they just get dataset audio from various sources and train it as "RVC voice model".
yeah some do
I think they need to have a softer voice if their voice is deep
I have to lowkey soften mine and it's hard as
bro same as my dad
dude thats what im sayin
he's in the other room working 😭
we cant get caught lackin
I need to be home alone
bro do you have yours set up?
i paid someone 100$ to make me a custom model without the voice acting part
i do
mf ran with my money
😭
wanna test it out?
no way
yeah
The audio quality of an RVC voice model has to happen with audio dataset, settings and how model maker(s) trained it.
send
you got scammed
yeah im trynna like make one or summin
dude i cant find a good model
i lost all my valorant kitten models cuz i switched ssd's
wait can I test mine out then
There are voice models in #1175430844685484042.
amanda silvera or smth
ty
ok spam ping me
YO
You can train a voice model by yourself, as long as your PC GPU good. 
warming up

thx
For conversation about "call 1" voice chat, go to built-in text channel of that voice channel.
okay daddy
same thing as June
The next month from June is July.
Yes
What are you looking for?
its laggy-
bro wydm
im tryna hear whatever his testing so i can see how it sounds like
come to call 2
@candid crater
im already in
i hear it
paid commissions are not allowed nor supported in this server, so it is your responsibility for getting scammed. but anyway I'd advise you against doing that.
how is it?
true
prom
Sometimes sturry and laggy
The funny thing is that some ramdom members would still join your main voice chat, even if you have an intention to test your voice changer, so a voice chat named "join to create" might be preferred for your case.
do yk how to fix it?
only max 2 can go there you tard

i dont know tbh im already tryin find settings and turn onn internet
I can't fix it myself because it's hard to playback and stuff
lowkey confusing
imma try the other one
wokoda or sum shi
im on tht its confusing
for Wokoda, it says all of them need to be in the same file before extracting
do I install everything on the github page and then get one common file then do that?
Namari?
like there are about 10 files there
Download https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001 and https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002. These two files have to be in the same folder, use WinRAR or 7-Zip to open the .zip.001 one.
plus 2 source code zip files
okay so I then create one folder and put those 2 zip folders in
will that folder also be a zip file?
What? Simply, .zip.001 and .zip.002 should be in the same folder, and that folder would be a "downloads" folder. It's not necessary always be "MMVCServerSIO" just for these two split zip files.
where do i find this ?
look up
here
download these 2
session option-
what?
NVIDIA GeForce RTX or AMD Radeon RX?
any idea why I get this when I try to run deiteris w okada fork?
im on lenovo-
This is b2332 W-Okada fork, not b2397 W-Okada fork.
Yes, but which GPU does your Lenovo PC has?
intell ( R )

Still following guide for Deiteris' W-Okada fork or something? The Tg Develop's W-Okada fork is a continuation of Deiteris' one that you might be overlooked.
You sure an integrated GPU can work with W-Okada voice changer? Sounds unlikely.
im new,..

what is that-
oh the dropbox
does tg develop’s w-okada have the same settings as the main one? like the “tune” and “chunk” settings?
Okay what do i do there i got account there
Tg Develop's W-Okada fork has pretty much the same settings as Deiteris' W-Okada fork, meanwhile the original W-Okada has different chunk and extra numbering.
Follow this guide. https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/
Last update: September 6, 2025
the thing is there is no internet Option.
is there crepe_tiny in tg develop's w-okada fork? @hallow thistle
There are only rmvpe and rmvpe_onnx in Tg Develop's W-Okada fork from start. Some other additional or legacy F0 models are available for optional download in "advanced settings".
what's the difference between Crepe Tiny (Onnx) and Crepe Tiny (Pytorch)?
Any model that says "onnx" is usually for non-NVIDIA GPU. "Crepe" models are legacy F0 models that often superseded by newer rmvpe.
what happened to this setting? I can't find it in tg develop's w-okada fork was it renamed or something?
These settings will show when you upload and select a voice model there, much like Deiteris W-Okada.
I don't see "Tune"
the thing is there is no internet Option.??
On your Kaggle notebook, do these.
getting this error when I press the start server button with tg develop's w-okada
No shows version and other stuff.
make sure you have done phone verification to enable that feature
i id
I did.
I'm now assuming I create a folder
and put them in
zip folder or normal folder?
you mean the "Internet On"? that's obviously as shown in namari's screenshot
or do I do this after opening .001 using 7-zip
Yes i verifed it.
does anyone know how you fix this? Pipeline is not initialized.
and you still don't see it? or have enabled it?
um what now?
I don't still see it
I see a folder called voice changer window, etc etc
expand all the dropdowns at the right pane pointed by namari
there are 3 files, 2 of them is force gpu clocks and the other a folder
mmvcserver
is it suspported by certain country?
do what I just said first
@hallow thistle do you happen to know how to fix this?
after clicking start server
Namari mate, so after extracting .001, both files were put in a folder and I went to a folder called MMVCServerSIO and it's an application, I double clicked on it and now it opened the cmd and is installing a bunch of stuff which I'm gonna wait now
I was busy trying to prepare an answer.
Can y'all please give me time to respond?
I mean expand the "Session options" first, and you should have seen it
There is no internet options or acceletor
@candid crater this^
On W-Okada voice changer, if audio mode is set to "server", your audio devices all set to "WASAPI" and selected a sample rate of 44100, the program would fail because of mismatched sample rate.
should i do client or server? I'm trying to run it locally on my pc
??
because I tried both client and server both same error
The supposed sample rate is commonly being 48000 Hz, because most integrated microphone (if uses integrated sound card) and "Line 1" from VAC Lite can support up to 48000 Hz, regardless of how high the sample rate of your speakers/headphones.
find "Copy & Edit" button
Go to https://www.kaggle.com/, see if there any notebook named "Kaggle_RealtimeVoiceCh", click on it, and you'll see something like in my second screenshot.
Your turn now.
I'm still getting 'Pipeline is not initialized.'
okay so I double clicked on that application and it installed the required things for it, now the app hasn't launched yet
this is in my downloads page
how do i know when the installation are done?\
it'll probably say
oh shoot
didnt edit
One Kaggle notebook wouldn't work itself without any human input.

To start a Kaggle session, go to "Run" menu, click on "Start session".
I also extracted 001 file and then put 001 and 002 in a folder
but I didn't do anything for 002
I'm more of providing settings and known fixes for a program, but if you ask me how to simply run the program, I am not supposed to tell you like that.
not working no matter what i do @hallow thistle rip
I legit installed the stuff you gave, extracted 001 and it created a big file with a bunch of stuff in it then I ran that exe file and it downloaded a bunch of things in cmd and then I put 001 and 002 in a common folder
How about your full screenshot?
how did you launch the app then?
that's the opposite, the 001 and 002 should be in the same folder before extracted
oh
cpu is red
wait so do I have to delete everything and redo it?
I simply double click on "MMVCServerSIO.exe", wait for several minutes and it launches in a web browser program as usual. I do things simple, I didn't have to complicate either.
I suppose so
okay and last question
you said they have to be in the same folder
do I extract them seperately or do I extract the folder?
like put them in, turn the folder into a zip file then extract it?
it can be at anywhere, e.g. D:\
but at the same path
for both files
I don't understand
Here's your settings:
Audio mode: server
Sample rate: 48000 Hz
Audio driver (or audio API): WASAPI
Input: [WASAPI] microphone
Output: [WASAPI] Line 1 (Virtual Audio Cable)
Monitor: [WASAPI] your speakers/headphones
Chunk: around 100 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 4090 Laptop
Pitch extraction: rmvpe


