Hey YouTube folks, how are you doing? Welcome back to another video here on the IAP channel. And attention, we have news for you now, and I want to show you how it works.
If you don't want to install, right, let me close my Ollama here that I activated. If you don't want to install our application via Pinokio to make your music, you can now install it without Pinokio. So, for those who already know how to install it, right, I'm going to show you here.
First, let me show you what the EPlay application is for. It does the following: it structures the prompts for your music. So, your music will be structured for you to use in Suno, and also in others, I forgot the name, Udio , uh, Music, sorry, Music, among others , okay?
So, it's an application that will certainly help you a lot. To structure your music. I want to show you how you can perform the installation.
There are two ways. You can download the ZIP file and you can clone the repository here. And that is what I will do just to show you.
Remember that to clone it, you must have Git installed on your computer. So you can search here if you don't have Git installed, okay? And then you will install the file, see, Git for Windows, do you see?
It's this first one here. You will install it, and it says here how to install it. Ah, on Windows.
Then you will come here and perform the installation on Windows. It's not difficult to do, okay? See, you will download it here, look.
Do you see? My 64-bit. What is your type of Windows?
You will download and install it. So, I'll even click here for you to see, look. It downloaded, do you see?
And then it will install. Afterwards, you come here and double-click to install. Got it?
It is simple and easy to do. And then, after that, okay? Once this installation is done, you come here, copy this URL, and go to the folder where you want to install it.
In my case, mine is already installed, it's right here. I did it last time using the zipped file. But it's simple, you will click here on the address bar up here, do you see?
On your Windows, look . You are going to click up here, type cmd, and hit enter. This window will open for you.
You will type here git uh , clone, OK? and you will put the URL. By right-clicking the mouse, you paste the URL.
Once done, hit enter, and it will start cloning the repository. Anderson, but you already have the repository there. That's right, that's why it did it so quickly, okay?
But it will clone the repository there. You can see here it cloned another one from the 11th. I won't use this one, right?
because otherwise I would have to download all the weights again. But for you who just downloaded it, you just come here, look, to start. Bat.
This is if you have Windows, if you have iOS, ah, Apple's, right? I don't know, I don't think it's called iOS. Ah, or also if you have that penguin one, which I forgot, Linux, you will click here to install, all right?
But uh, I will click on my other one that is already installed. It will install several things. You will allow the installation.
I will delete this one, right, since it's light. And this one here, you can already see in its properties, look there, five. It's going to be about 10 to 15, I think 10 teras, look.
11. 11 teras there, because it downloaded so I could generate the music right on my computer . Ah, Anderson, but I don't want to generate music, it gets too heavy.
You just come here, you will take the uh checkpoints part, see? CPTS, checkpoints, right? And you will remove this here, look.
You will remove this, you will remove that, you can remove all of these, it won't make a difference for you, okay? Why? Because this is where you generate the music.
Ah, but I don't want to generate music. So you will delete all of this here, look. If you check its properties, you will see there are, look, five teras just here, okay?
Ah, are there other things that can be deleted? Yes. You can delete other things here.
It's just that I already have several generated songs here, right? Some generated songs there. And there is also uploads.
Yeah, it didn't do it, the uploads didn't appear here, but there is quite a lot that I've already done here. Engine will also be here in models, I think it's models. So here there are some things I will be using in the future to generate, uh, for us to generate images , generate videos.
So I am preparing this, using Minimax. So you can see that it already downloaded some things to have them ready for the near future. So do you need to use this?
No, you don't need to use it now. If you want to and you delete it here, there won't be any problem at all, okay? So you see there is a lot of stuff here, look, that is already preparing for us to generate something, uh, really cool for you here in the future.
So that's it, okay? Removed this here, it will already remove a lot of stuff. Now let's go.
Eh, I think here, I think here it's just this, let me see. It's not its index. I think here is only what you will remove.
You can see that the other files here are lighter. If you remove here, uh, the checkpoints and remove here the the others, right, that I showed you here, in the server there is nothing you can delete. Public neither.
Just for us to see. It's really just in the engine. Here there will be the checkpoints again.
So it's as if it were twice, right? You downloaded it, removed it here, removed it here, there won't be a problem, okay ? You can set it to run.
Now I will click here for us to start. I noticed that when you do it through Pinocchio, it starts faster and it generates things faster, okay? But when you do it without Pinocchio, it doesn't do it as fast, right?
You can see that it takes a while to start. It takes a while because it has all the checkpoints to start. Let me even delete here.
Not delete, uh, turn off here the the antigravity, which I use antigravity to do the upload, to do the posting here. It accelerates this whole process for me after I mess with the application's structure, right? Here you will have, look, installation, how to start, main features, visual tour, screenshot, synthesis, uh, how it works, you will have everything explained here in Portuguese, okay?
So, take a look, see what you think, okay? See what you think, if it works for you or not. Now it's here, look, it is already installed.
When you generate a song, for example, a musical style here, uh, I want a rap. When you go to generate another one with the same lyrics, you have to come here to new, okay? Look, you're in here, you'll have to go back and return to the new one, okay?
Uh, just a quick explanation so you don't keep trying to regenerate here, otherwise it will mix everything up, all right? Then it will mix an R with what you, for example, right? You did an R, then you want a rock, it will mix everything together.
Uh, well, here, look, you go to API key. The first thing you have to do after installing the Ollama API key is, you'll come here and download Gemma 2. For me, it's the best.
You'll copy it, okay? Once you've copied the command, you'll come here, uh, and go to PowerShell. Do you see, look?
You just type Power, and this will appear for you on your Windows, okay? And you'll just copy it here, look, right-click and press enter. It will install Gemma 2 9B there for you.
Anderson, but I don't have Ollama. Install Ollama. Look, it's this application here, okay?
You need to install it. Maybe you don't have it. You'll type into your Google, look, ollama.
Why this? If you want to use it without needing an API. If not, ah, Anderson, I don't want to use it, to use Ollama.
No, I don't want to. It has all of this here that you can download directly to your computer, which are, ah, light, they aren't heavy, okay? But this one here is the one, it's actually even the recommended one, because it's creative, ultra creative, and it's quite light.
You come here and enter, for example, the API key, uh, from Google, from Groq. Groq with a Q, not Elon Musk's Grok, okay? Ah, this one here.
So you can, look, Nvidia. Ah, here in this case it's Ollama, right? Nvidia's default model, because I downloaded, I put my Nvidia API here, okay?
Create there, you can create your free key and there are several models. Look at how many models there are, right? So I can, for example, put here Google Gemma, Google Gemma 3 12B, which is even better than this one, look, Gemma 2 9B.
I can leave it selected here. And there are other API keys, such as for example OpenAI, which is also ChatGPT. ChatGPT now has a free API.
You just have to come here, look. Chat, uh, GPT Free APIs. You come here, look.
And then you can get it here, see ? Do you see? You can go to it and search for APIs there.
Ah, let me, I just don't remember exactly, but take a look around. Ah, how did I do it? I think I even saved it here in the AIs.
I think I even saved the link here. Let me just check. And the paid ones here, look.
I saved the link. It's right here , see. Plat, do you see?
Platform. Open. open.
com APIs. Here is the API I created. You will come here and create your free API.
You didn't see this one, right? It’s, it’s covered up there, the API. Uh, because I'm using it, so I can use it for free there in our application as well.
So, it's up to you . Ah, I want to download it, I want to use everything on my computer, have everything uh, all there on my computer and not generate anything outside of it . You can download Ollama.
You will type here Ollama, do you see? Ollama. com.
Come here to download and download it, right? And then you download the model in the way I told you. You will take the, look, you will download here the, uh, do the download for Ollama, right?
You can come here for Windows, for example, and it’s the same thing. How do I perform the installation, Anderson? The same way, right, that I showed there, you go to PowerShell and paste, right, this this this code here.
But you can click here, look, on manual download, do you see? And that's it. You come here, look.
After it downloads, you click and install it normally. Then you will do what I said, you will take this, copy this code and throw it into PowerShell, right, for it to install for you. Got it?
Uh, well, I think it was understandable, right? It was possible to understand what we talked about. Once that is done, installed in this way, this also applies to you who will install via Pinocchio there.
I already showed how to do it in the last video, do you see? Here there are even the sites for you to get the free keys, look. There are the sites for you to get the free keys there, okay?
Another thing I didn't tell you, I want to talk about quickly, is this issue of identity. You can put who you are here so that the AI, if you're using it on your computer, eh, I'm going to remove this in the next update. I forgot to remove this, it's not needed because it doesn't have a password.
Ah, I will, you write here exactly, right, what you like most and it will learn from you to generate something you like better. Got it? Going back here to my projects in Promptimestre, I already explained to you in the in the other in the other video, right?
These are the musical styles, these are, and you can add here , these are the prompts that stay in its engine there for you to use, okay? And these are the prompts I use, okay? Now I'm going to come here, look, new project, look.
It's here locally. I can come here to Nvidia, for example. And in Nvidia, look, it's already using it there.
Ah, I can put Gemini here, see? And here in Nvidia I have that that key , right? I I didn't save it.
Now I see I need to get the key saved here. I saved it in Pinocchio. But anyway, I could use it.
Just put the key here and save and that's it, right? And then I have all of these here, look. ChatGPT.
Let me see, look. Chat is here, see? And I can use it.
So, let's use it just so you can see. I'm going to create a happy song, right? I'm going to create here, look, generate lyrics.
Generate a gospel song. I know there are people who don't like it, but oh well, it's what I listen to, okay? Everyone has their own musical style based on Psalms 23.
You can do it in any style. I'm going to click confirm. It will generate via the ChatGPT API and we will continue here.
Logically, since it's a free API, it will take a little longer than if you used Ollama which is internal, right? So I can have the option to use it there. I'll let it generate here, I'll open another tab.
If I use Ollama here, even though it's already generating another one here, I can put Ollama here, right? and it will generate faster for me, okay? Let it generate here.
When it finishes generating, then we, look, it's already finished. So, it's here, look. It left the language, built it here, right?
And now that it has built the lyrics for me , I will state the style I want. I want rock. Oh, and I can just type here and it will search for me, see?
But here I already have metal. OK. Done.
I'll just put two. Here I set it to high quality. I open the creative arsenal.
And here there are other new features. I want a male voice. I want a raspy and hoarse voice.
So you already set the vocal style you want here. Here are the instruments you want to add, but if you don't put anything, it will add the instruments it wants, okay? So I can add guitar, I can add, for example, look, acoustic guitar, which is the electric guitar, electric, guitar.
This will be here because the style already includes it. Here is the mastering, how you want to improve it, right? So, for example, I can set it here, which is ideal for rock.
Oh, here too, look. And for the atmosphere, I can set it to live stereo, but I won't put anything there. Here, look, distortion and overdrive, which for rock, that's important, right?
And here I have the saved presets in this other tab. I will click here to apply. It applied it, see ?
And now, after it applied this, I will generate the final prompt. I click there on generate the final prompt. Remembering that it is generating with the ChatGPT API for free, okay?
So, it will generate it and I'll be back in a little bit. Hmm, there was almost no time, right? I pressed pause, and it already generated it for me.
So, it's right here, generated prompt, right? Global prompt, uh, the musical style, lyrics, everything structured, right? Ah, here for me.
Chorus, everything is structured here. The style is already here. And if I just want to throw it into Suno, I'll just click copy here and throw it over there.
And here it's the same thing, copy and I throw it over there, okay? Oh, but let's say I want to generate it here in our own app . If I downloaded it the way I downloaded it here without and didn't delete any file, I will click generate here.
I will set it up to 7 minutes, but it won't make 7 minutes, okay? I'll make that very clear to you right now. Ideally, if you use Ollama, it will, that specific Ollama AI style is generating this structure much better.
So I'll even generate it here again, just so you can see what the difference is with Ollama. It generates it and leaves a note in the description here that it is in Brazilian Portuguese. I don't know why sometimes by switching the AI it doesn't include that, even though it is one of the prompt rules.
So, depending on the one you choose, it doesn't obey, see, because it hadn't done it. So I would have to take these lyrics again, and do what I told you, look. You come here, you create it again, you paste the lyrics, oops, Ctrl + V.
Then you put everything in here again. Rock. Uh, where's the IN rock?
Not here, metal, right? Metal. Then you're going to put voice here.
Oops, male voice. I didn't put anything here. Mastering is this one here.
The rhythm is this one here. I didn't put anything for atmosphere. and distortion effect.
Done, applied. I'll generate the prompt again here. You'll see it will be faster because it's generating.
Let me close this so there aren't too many AIs running on my computer. It's generating here directly in Ollama. I don't think there's much difference in the generation either.
Yeah, it didn't generate it here. I don't know why, but I'll fix it later, I'll see why it's not putting that it's Brazilian Portuguese in the global prompt. I want it to include that so the AI sings in Brazilian Portuguese and doesn't cause issues.
But let's go. Well, it put it, I'll go back here, look, 7 minutes and let's generate it this way. If anything , uh, you put here, look, Portuguese, uh, Brazilian Portuguese, right?
Brazilian. Brazilian. I don't remember if it's spelled with an N or an M in English.
I think it's with an M. Let me check here, guys. I don't remember anymore.
My mind just went blank on the English. Oh, oops, Brazilian Portuguese . It's Brazilian with an N.
So, I'll come here, look, put it here, put a comma. Done, I'll solve the general problem, because it's in the general prompt. And then you see it arrived here, look.
He is my shepherd, my protector. And here is the last part. I can take a song and use it here as a base.
I can make a cover of these songs and everything else. I'll leave it here at 32, but you can go up to 64. The higher it is, the better the quality.
To generate faster, I'll set it here to 32 and click here to generate the song. In this first generation, it will load to the GPU. I think it will take a little longer because I'm recording.
So I'll pause, wait for it to generate the whole song. Then I'll bring the result here for you. It has already started generating.
While it generates, I want to show you another function here that I almost forgot. After you have generated the final prompt here, only after you generated it. If you haven't generated it, I'm going to touch the microphone and make noise, guys.
If you haven't generated it, this won't appear . If you have generated it, you will click on session rules. Do you see here ?
Session rules. And this is the secret trick of this new version of our EAPlay . Uh, it's generating over there, look.
Here you can change, add melismas, right? Uh, here you already have an edited prompt. So, here, for example, ah, I want to add a melisma here.
Let's go here, look, melismas, right? Ah, I want to add a melisma. I want to put it on the word.
Then I'll take it here in the pre-chorus and lead me here, on the word. Then I'll put here, a melisma on "preciso estar". Yeah, it doesn't select two.
You have to do it one by one, insert into the word. Then you come to the word again. Oops.
On the word "estar" here, extend melisma. " Estar". After "estar", you come, ah, to "preciso", see?
So you put it, look, on both "preciso", "estar", "estar" and " preciso". It doesn't matter. Raspy voice at the end.
You can add it here. These are just a few, but here you have more, okay? You have more functions here, look.
Long notes. Ah, you can put here, look, vibrato. Ah, I want a vibrato here, uh, in the header, at the end of the line.
I also want a vibrato after the melisma, right? So, look, I placed a vibrato here at the end of the line and then you click save, see, rules saved. And when you come to generate, it will already be different, right?
Ah, I'll set it here as if I were going to generate, but I can't generate now. And you will see that, um , when we get here, look, look there, it's already here. See?
And you can do this throughout the lyrics, the whole lyrics, okay? You can do that. This is very cool.
You can edit the prompt to make it generate a really cool, very distinct song, with your, uh, your final touches. Anderson, but I don't want to generate it inside the app, that's no problem. You do the same thing I showed you here.
Right? Come here, look, the final prompt is already here, look, it's already here. You're going to copy it, right?
And paste it into Suno. You'll paste it. Suno follows instructions very well.
It's a shame they're shooting themselves in the foot by doing what they're doing, but Mureca is also very good at following this here. And the Mureca audio is also very cool, okay? So you can use this app to run your test.
Oh, I liked the song. Use the song as a base in those others there to make a cover and improve it, right? It's a solution so you don't keep spending credits there if you have a good computer.
But I tell you that here we can generate audio with high quality, really high quality. So, after it finishes generating I'll be back to show how it turned out. Done, the audio generated here, it took a little while.
Ah, about two minutes here. Actually, I think it says, right? It doesn't say when it's generating.
Oh, no. It says here, look. 1:39 PM and now it's 1:40 PM.
So, it's been 1 minute since it generated. We just have to take it. It'll take, I don't know, about 5 minutes, but if you do it through Pinóco, it generates faster, okay?
I don't know why that is. Guys, when you generate through here, it already has a humanizer. It removes the AI watermark.
When you generate with this little button pressed, which is the case, it already did this, what does it do? It does AI anti-detection and it does mastering. Everything that is below what our ear, our ear can, can hear ah from 20 eh Jesus, I got lost now.
If I'm not mistaken, it's from 20 to 20,000 Hz, right? Ah, up to 40,000 Hz. If I'm not mistaken, okay?
But I know that AI generates audio below what the human ear can pick up. So, when we listen to the music, we don't understand, we don't see, we don't hear it. People also use this for subliminal messages.
So for those who know about that, well, you already understand. And AI generates this audio below those 20,000 Hz, right? And this application that we created removes this audio so that YouTube and other platforms cannot identify that it was generated by AI, okay?
I think this is important. So, you can—if you generated it without pressing this or thought it was still there, you click here; it will generate a new audio, the same audio, but it will actually master it, remove it, and leave it as is. You'll see that, sometimes, there will be a small slip, like any human who sings happens to have in one song or another.
We who just listen, not being musicians, sometimes don't notice, but those with a musical ear know that not all songs turn out perfect, right? Especially when recorded. People tweak things a lot to improve that.
But let's go, let's listen here. This is the result. Folks, I thought it turned out phenomenal.
It's good. Phenomenal. It generates in wave, 48 kHz and in stereo .
OK. Look, it's already saying here, look. Humanized audio, metadata, ultrasonic cut, and tape warmth.
That's what it added. It removed the AI information and added this here to make it humanized. The peace of the Lord guides me.
In green pastures, He leads me. Calm waters renew me. He takes me along paths of light.
Have no fear, have no doubt, for divine love strengthens me. His hand holds me with care and leads me to where I need to be . He is my shepherd, my protector.
In His shadow, I find shelter. He leads me to a full life. In His presence, my heart vibrates on days of darkness.
>> Guys, look at the quality of this audio , right? Look at the quality of this audio here. This would be another style of audio, already tweaking, right, in the adding melismas, you know, adding some effects there, okay?
In case it happens that, oh, it sang in Portuguese from Portugal, kind of mixed it up, just put it in the global prompt, right ? You can do this before sending it to Suno, you put it here, Brazilian Portuguese. OK?
And then it will sing in Brazilian Portuguese, alright? This applies to anything you’re going to generate for music, okay? If that happens, you put it in brackets, in parentheses, right?
Ah, not parentheses , my God, I forgot the name of this thing. I think it's square brackets. No , brackets are that, that other one.
I forgot. You put it here, look, in this file right here. Oh Jesus, have mercy on me.
And you put Brazil in Portuguese and it will sing in Brazilian Portuguese, alright? I hope you enjoyed it. We made this application with a lot of care for you, and you can help us by hyping this video, leaving a like, sharing it, and also when you go to download the application, which will be there in the video description for you to download directly to your computer or use Pinokio to download it, which is faster, right?
right? And Pinokio installs everything you need, because suddenly there might be things not installed and that will cause problems. It’s no use complaining about the application, it’s things that are missing on your computer.
Eh, and Pinokio already installs it, right? And every computer, every situation is a different situation, right? And Pinokio already does this installation for you.
So, there are these two ways. If you already understand a bit more about computers, this will be easy for you to do if you don't want to use Pinokio on your computer. That’s it.
So, dear friends, I hope you enjoyed it. Big hug , God bless you, and until the next video.
