Thank you for watching! If you are trying out the program and have some issues, please open up an issue over at github instead of here at youtube. github.com/Koischizo/AI-Vtuber/issues Thanks.
This is a phenomenal video, and has inspired me to add on to this AI and really make it my own! The only problem is I have never coded in my life, and have no idea where to start. Any advice on a good place to start would be appreciated :)
Hello, I don't want to use eleven labs, so I want to modify the pyttx3 voice to Spanish but I don't know how to make this change, I would appreciate your help
Safe to say , our flesh tubers with extra steps are gonna be here for a little while longer . Banger tutorial video though , learnt some things and had a laugh too XD Edit : I called it in the last video , I knew you were a undercover Capipalist
I think another cool idea would be to make it a desktop application, like just having it on your computer so you can talk to it and it can find stuff or open stuff for you but thats another step that i dont even know if its possible to do, just like the MAID from Sakurasou if u know what i mean
You just took the words of my mouth, wanted to create such an assistant recently and I don't see why it shouldn't be possible at current moment from the technical point of view. Using voice commands to do things on pc like opening some programs is not all that difficult. With gpt 4 it would be even possible to make screenshots of something and let the assistant analyze them for you by taking it from your clipboard.
I mean, you could try utilizing Pygmalion A.I, to say, circumvent some restrictions. Or just, increase the degree of customization for your A.I Vtuber.
This would be so helpful as I do have plans to create my own AI assistance for my live streams as well as handling workloads that require critical thinking and decision making plus more if it turns out this good at doing so.
I was thinking about doing this now that I find this will make everything way easier I don’t have to do hardly any code between this and other RUclips videos
This is amazing, but costly tho....in my country I cant even pay for the OAI text model unless I use a VPN witch translates to pay even more, because they simply don't have the option to let you pay...great project tho, i guess we have to wait for other free options
Great work on this, man I might want to try making some. But for me it's going to be a companion while I'm playing a game. Like the AI will comment on me and youtube chat. Keep those tutorial going, dude!
We have more things now that can improve this. What about the new auto gpt? It could be used to improve the results even faster every day :) Quite frankly I can't wait to try this for myself and see how it goes.
it wouldn't be particularly hard to do this. Infact all you really have to do is to replace the youtube chat input with a text transcription input of your voice (use whisper locally or any of the cloud services for it) and maybe change the initial prompt a bit to make it somewhat less streamer and more helpful in terms of personality.
im generally not a fan of AI, but love AI vtubers that are unhinged to the point were theyre both complaining about NGE and praising it for the same reasons. Like, perfect accuracy AI is a waste of tech when you can just make them into legit gremlins.
I appreciate the transparency to share your process on this interesting topic and even open a git repo. After deciding to go down the rabbit hole recently, I was disappointed when another guy seemed very hush hush about how he went about making his.
If you're talking about Vedal, it kinda makes sense for him to not open source his project since it's making him money I mean, people are probably gonna figure everything out eventually, let the guy enjoy his monopoly while he can, I think he earned it
Nice to see someone open-sourced an AI vtuber (unlike those another dude who uses Whisper and GPT-3 for his own AI). Not that it matters since I decide to follow yours instead. Cheers!
Thank you for making it possible to create my dream waifu! :) Since I'm a complete newb, I hope you can clarify a few things about this project. 1) I've executed the file using the second livestream link to test, and I immediately noticed the A.I.'s responses are no where near the same. Is this because the A.I. still needs to be trained? If so, can you recommend a resource and instructions how to implement the trained data into this project? 2) The TTS output is "very" rough. Is this purely a potato PC issue? Thanks ShizoDev for all your hard work. (づ ◕‿◕ )づ (づ ◕‿◕ )づ
simple amazing, I have all set, I just have a doubt, where did you paste the converted to python code? It's giving me an error in line 15, the EL_voice line, I did something wrong for sure, thanks a lot, let's the real vtubers revolution begin!!!!!!
So um... i tried it. And it actually works! ... for the chat part. When it tries to execute tts() function I got: "PermissionError: [Errno 13] Permission denied: 'C:\\Users\\CD86~1\\AppData\\Local\\Temp\\tmpy5oeix9k.wav" Any collegues know how to fix that?
Thanks a lot man. btw Can u tell me which virtual camera you were using to move her head,streaming other stuff for A.I. vtuber streaer set? And can you give me that link of the repository? Would really appreciate your help
im one year late to the competition but imma go ahead and program myself a 2B who will respond (hopefully) to my voice so I finally have someone to talk to. Im hobbyless
So modified the code should monitor both RUclips API and twitch API I've also added expressions to show emotions the only thing I'm having trouble with is the 11 Labs API I can't get her to respond correctly with the TTS
Im having one trouble man my program is working totally fine just somewhere have problem in audio or maybe im not good at doing settings so its actually generating text but not talking and not moving lips also and pls tell which software you are using for streaming ?
This is amazing,but can I ask a question?Is it counts the TTS's word number as I'm using AI?Because it has a limit amount right?I'm using free now,so I'm worried.
Thank you for watching!
If you are trying out the program and have some issues, please open up an issue over at github instead of here at youtube. github.com/Koischizo/AI-Vtuber/issues
Thanks.
This is a phenomenal video, and has inspired me to add on to this AI and really make it my own! The only problem is I have never coded in my life, and have no idea where to start. Any advice on a good place to start would be appreciated :)
looks liek a great start, kinda wanna try llama tho, i'll probably write a similar solution to yours
how'd you get this on a chrome book lol im too broke to even afford a 1 cent computer
@@ImPossiblyElgia hey, did you progress through the github stuff? I'm also in the same boat haha
Hello, I don't want to use eleven labs, so I want to modify the pyttx3 voice to Spanish but I don't know how to make this change, I would appreciate your help
This madlad is making AI waifus real
I would join his team and let him be my mentor/teammate. It would be nice to work with him
thank you for your kindness ,sir
"Touhou is space invaders on crack."
I have no objection.
The first depressed and toxic AI vtuber. Technology sure is amazing.
then Neuro-sama took that XD
Safe to say , our flesh tubers with extra steps are gonna be here for a little while longer . Banger tutorial video though , learnt some things and had a laugh too XD
Edit : I called it in the last video , I knew you were a undercover Capipalist
Tell that to Neuro
@@jeffbrownstain BRO XD
I think another cool idea would be to make it a desktop application, like just having it on your computer so you can talk to it and it can find stuff or open stuff for you but thats another step that i dont even know if its possible to do, just like the MAID from Sakurasou if u know what i mean
that'd be awesome
ChatGPT can already search info for you and even more, it doesn't sound impossible having jarvis-like ai desktop assistants in the near future
@@veto_5762 jarvis, show me this guy's balls
You just took the words of my mouth, wanted to create such an assistant recently and I don't see why it shouldn't be possible at current moment from the technical point of view. Using voice commands to do things on pc like opening some programs is not all that difficult. With gpt 4 it would be even possible to make screenshots of something and let the assistant analyze them for you by taking it from your clipboard.
Jeez...
The era of *true* virtual youtubers is looming, guys.
Let's simp onto *programs* .
Yo this voice goes crazy too it’s so realistic
This is exactly what I was looking for. An AITuber tutorial with an open source repository and a schizo loli. Perfect!
Congratulations, hope you got interviews by vice,verge and thewired. Btw gpt-3 and elvenvoice can replace by anything as long it work
U just opened my eyes today.. I'M CREATING MY OWN AI VTUBER TODAY!
I am trying your projects man. Can’t wait to see the final results
thank you for this, hope everyone can make their ai waifu soon
Dam its you hi lol
@@natsudragneelthefiredragon yo
Every time I try I get Error, I hope the issue isn't AMD GPU,
But where can I get this Koishi Model?
I love how to stopped "borrowing" towards the end of the project.
Just what I was needing like 2 days ago. You just won me man. Keep the hard work!
She is absolutely precious, I love her!
I mean, you could try utilizing Pygmalion A.I, to say, circumvent some restrictions.
Or just, increase the degree of customization for your A.I Vtuber.
amazing.. this will be nice to learn some new stuff. Thank you !
OMG; thanks a whole lot for sharing implementation details
This would be so helpful as I do have plans to create my own AI assistance for my live streams as well as handling workloads that require critical thinking and decision making plus more if it turns out this good at doing so.
Nice video !! You give a dream to build the own AI Vtuber challenge ,Thx you I will try my best。
Thank you, hilarious. I shall do this
can't stop focusing on the COM3D2 music lmao
especially when the fucking music came on at 8:18 holy shit
How does this man ONLY HAVE 4.28k subs
God there's so much fucking coding theft, I love it. Absolutely unhinged, schizophrenia/10
A hero we need but don't deserve.
I was *ahem* inspired by your code.
I like her voice lowkey.
What a legend, thank you sir.
amazing! thanks for sharing your knowledge sir!
Yea but he sadly didn’t say where we can get the Koishi Fumo Model
I was thinking about doing this now that I find this will make everything way easier I don’t have to do hardly any code between this and other RUclips videos
thank you so much schizophrenic developer!
Ima gonna borrow your code.. and crank it to 11. Thank you vermuch for sharing
Me without knowledge of programming and limited english and free time: ok... I'll try
This is amazing, but costly tho....in my country I cant even pay for the OAI text model unless I use a VPN witch translates to pay even more, because they simply don't have the option to let you pay...great project tho, i guess we have to wait for other free options
The tor browser is a free vpn, you just have to get lucky to end up in a country you can pay in.
A very fine choice of music.
I approb.
we need more, maybe time for insane
Great work on this, man
I might want to try making some. But for me it's going to be a companion while I'm playing a game. Like the AI will comment on me and youtube chat.
Keep those tutorial going, dude!
Hey man!
I thought about the same thing!
You did it? It worked out? Any tips?
just found this youtuber and i instantly will be subscribing
The Koishi made me click on this at lightspeed
We have more things now that can improve this. What about the new auto gpt? It could be used to improve the results even faster every day :) Quite frankly I can't wait to try this for myself and see how it goes.
I thought we would need at least 1 more year
Thank you comrade.
Could you make a version that’s more of a virtual assistant rather than streamer? It would be Siri 2.0 😂
The Japanese already made one called Gatebox
it wouldn't be particularly hard to do this. Infact all you really have to do is to replace the youtube chat input with a text transcription input of your voice (use whisper locally or any of the cloud services for it) and maybe change the initial prompt a bit to make it somewhat less streamer and more helpful in terms of personality.
that ending man, I still miss beatani
So cool
This is really cool
Creating my own gf assistant but replacing the AI bit with prewritten responses to better optimize and run locally so no glowies can spy on us 🤭
I failed. OK I'm kms.
RUclips's algorithm is powerful
I'd love to be able to create one myself! But i have 0 coding skills, and it puts me off slightly...
Thanks, I am a beginner, learned a lot from your video
Hi sir I'm sorry i didn't see this video, thank you for the source ❤️❤️❤️
Teach me, Master, Sensei, Dean, Professor, Teacher,
koishi is good and good and very good
koish moo moo
AI breaking there filters
im generally not a fan of AI, but love AI vtubers that are unhinged to the point were theyre both complaining about NGE and praising it for the same reasons. Like, perfect accuracy AI is a waste of tech when you can just make them into legit gremlins.
I appreciate the transparency to share your process on this interesting topic and even open a git repo. After deciding to go down the rabbit hole recently, I was disappointed when another guy seemed very hush hush about how he went about making his.
If you're talking about Vedal, it kinda makes sense for him to not open source his project since it's making him money
I mean, people are probably gonna figure everything out eventually, let the guy enjoy his monopoly while he can, I think he earned it
@@john.darksoul i reckon that one day he'll open source it, maybe if he gets permabanned from twitch lmao
@@john.darksoul He deserves the monopoly. He's such a gift to all of us. I love that man.
@@john.darksoul He would literally make more money by letting people help make Neuro better 😂
NO NOT KOISHI SHE GONNA COMMIT WAR CRIMES ON LIVESTREAM
insert uh KKHTA here or smth
That’s a vary Heart Throbbing clam
so is there a twitch version of this would love it
Nice to see someone open-sourced an AI vtuber (unlike those another dude who uses Whisper and GPT-3 for his own AI). Not that it matters since I decide to follow yours instead. Cheers!
@SchizoDev You're the reason I'm learning python.
Thank you for making it possible to create my dream waifu! :) Since I'm a complete newb, I hope you can clarify a few things about this project.
1) I've executed the file using the second livestream link to test, and I immediately noticed the A.I.'s responses are no where near the same.
Is this because the A.I. still needs to be trained? If so, can you recommend a resource and instructions how to implement the trained data into this project?
2) The TTS output is "very" rough. Is this purely a potato PC issue?
Thanks ShizoDev for all your hard work. (づ ◕‿◕ )づ (づ ◕‿◕ )づ
Do you tts function works first time?
Mine is kinda stuck on "Permission denied" error while trying to load audio from responce
@@angryspearmen3614 Same error pls tell me if u could fix it!!
@@angryspearmen3614 Yeah i saw this already, it seemed more in depth but also harder, plus the google voice is kinda meh. Did this work for u??
@@timtvb7270 Yes. But i used different voice site (I needed foreign language :D). It took some time
@@angryspearmen3614 Wow thats sweet man! What tts did u use? Voicevox?
a step closer to future
That is my daughter. I'm adopting her, nobody else can, end of story 🥺🥺🥺 ok?
Vary Heart Throbbing if you
*V-Tubers are just FleshTubers With Extra Steps*
There 2 vary different things the only similarity’s is you need a Webcam
That’s it
-9842 social credits, your execution is tommorow
This is a nice concept. I just have a few questions?
1: Can this work with Twitch chat?
2: Why you don't have many subscribers?
you are my new hero!!!!!!!!!!!!
simple amazing, I have all set, I just have a doubt, where did you paste the converted to python code? It's giving me an error in line 15, the EL_voice line, I did something wrong for sure, thanks a lot, let's the real vtubers revolution begin!!!!!!
Can she Run with PygmalionAI ? but that would probably make her cursed .
Hey that's a GPD WIN MAX! I have one of those!!
I've never seen a laptop with analogue sticks and a d-pad
Hold up...
I hear a certain tunes...
one does not belong in SFW video
Also, don't abuse your meidos
"I have a date with my girls"
Oh no...
But also thank you! That's really inspiering
So um... i tried it. And it actually works! ... for the chat part. When it tries to execute tts() function I got:
"PermissionError: [Errno 13] Permission denied: 'C:\\Users\\CD86~1\\AppData\\Local\\Temp\\tmpy5oeix9k.wav"
Any collegues know how to fix that?
Just rewatched the video and saw that this part of code was written with chatGPT. Oh no...
Should be fixed, git pull and then reinstall the dependencies from the requirements.txt
i love ur content
She sounds way more natural than Neuro
Ikr
Would it be possible to make it run locally without services that require internet? Would it also be possible to give her computer vision?
That's so cool
Thanks a lot man. btw Can u tell me which virtual camera you were using to move her head,streaming other stuff for A.I. vtuber streaer set? And can you give me that link of the repository? Would really appreciate your help
Neurosama collab when?
I am in da video!
amazing
im one year late to the competition but imma go ahead and program myself a 2B who will respond (hopefully) to my voice so I finally have someone to talk to. Im hobbyless
fleshtubers with extra steps lol
There 2 things with the only similarities is that you need a Webcam
That’s if
97% CPU, I can feel it bro😔😔😔
So modified the code should monitor both RUclips API and twitch API I've also added expressions to show emotions the only thing I'm having trouble with is the 11 Labs API I can't get her to respond correctly with the TTS
Where did you add the twitch api and token in the code? Tried for both this one and the awdii ai but nothing works so far.
It's under config line 8 between nickname and user. It says token
@@evieyorisou2586 Weird I cant seem to figure out where to add twitch token and api requests without the pytchat constantly requesting an ID.
@@bel5578 You ever get the Twitch part to work? I've been searching all day
Thanks, its possible to use GPT 3.5 or 4? Or use custom voices?
Im having one trouble man my program is working totally fine just somewhere have problem in audio or maybe im not good at doing settings so its actually generating text but not talking and not moving lips also and pls tell which software you are using for streaming ?
ily
1:36 Gura: bruhhh
ara ara
Can you leave a link for the VTuber Model you use?
I don’t know where to get it as it looks really cute
is that a cm3d2 soundtrack
This is amazing,but can I ask a question?Is it counts the TTS's word number as I'm using AI?Because it has a limit amount right?I'm using free now,so I'm worried.
no way, koishi vtuber
Thanks for the effort, I'll try to see if I can run the ai, I only need to install the programs that notepad says?
Neuro-sama 2.0
I need it