what is going on guys? Welcome back. In this video today, we're going to learn how to easily create an image generation AI Discord bot in Python. So let us get right into it. [Music] Not all right, so a quick demonstration and preview up front. This is going to be our final result: a Discord bot which either reacts to messages that you send to it directly, or it also listens to channels, to messages in channels which it has access to. And what we can do now is we can type something like exclamation mark gen, and then for example horse, and this is going to generate an image for the prompt horse. Takes some time, especially because I'm recording now, and it's going to result in the image of a horse. Then I can do something like exclamation mark gen blue horse, for example. Let's see if that works. And in my case now, I'm running stable diffusion 1.5 here, so you can also use better models if you have the capacity to do so. In this case, I get this here. Okay, and then let's do

something else like maybe, uh, man on moon eating lunch. Let's see if that is successful, uh, but that is basically the idea. You can text it directly or it can react to messages, uh, yeah. Okay, this is not very good, but yeah, you get the idea. This is what we're going to build in this video today. All right, now for this project, we're going to need two things. First of all, a Discord bot. Second of all, an image generation model that the bot can use. And then we're going to combine everything into a single Python script which operates the bot. So, the first thing we want to do is we want to go to our web browser and visit discord.com/developers. Now, once you're logged in, you're going to see this page here, and here we're going to just click on new application to create our new, uh, Discord bot. I'm going to call this image generation bot. I'm going to accept the terms of service, create, and here now we're going to go to OAuth2. We're going to click down here on bot.

and we're going to grant certain permissions. For example, send messages obviously because we need to send images, attach files obviously because we need to send images, and then also read message history, view channels, and I think embed links can also be useful depending on what we want to do with a bot. But these are the permissions that we need. So what we're going to do is we're going to copy the link down here. This is going to be the invite link for our server. We're going to open up a new tab. I'm going to copy-paste this. And now I can allow, as an admin of the server, as an admin of the Neural N Community server, which you can by the way join as well. I will click on continue and authorize the bot to join the server with the permissions. Now I need to authenticate myself here with a code, so I'm going to do that quickly like this, confirm, and I'm going to also confirm that I'm human. There you go, invite the bot to the server. So the server, the bot is now part

of the server. I can see that this is the case by going to the welcome lounge, and you can see that here the image generation bot just slid into the server. So, I invited the bot, and the bot is now part of the server. However, the bot is offline because it's not operated yet. I didn't connect to it. In order to do that, I need to get a token. So, I'm going to go to the bot page, and here I'm going to scroll up here, or I'm going to go to the top, and I'm going to go to reset token. So, I'm going to do yes, do it, and I'm going to copy this thing here. Now, you can always reset this, but you're only going to be showed this token once, so copy it. Go to a file in your Python development environment, create it, call it token or whatever, copy-paste this here, and now you have it. And this is what you need to connect to your bot. So, on the bot page, we're going to do a second thing. We're going to grant

the privilege here, uh, which is the message content intent. The reason we need that is because the bot, of course, has to be able to read the messages that it sees. So, I'm going to enable this, save changes, and, uh, this now allows the bot to also view message content. So, this is what we need for the setup here. The second thing that we need is we need an image generation model. In my case here, I'm going to use stable diffusion version 1.5, uh, from Hugging Face. You can use any other version of stable diffusion. You can use any other image generation model, but you need to have some way to get a prompt, feed it to the model, and get an image in return. This is how you do it with this package. You just import from the fusers, uh, stable diffusion pipeline. You also import PyTorch. You specify the model. You create a pipeline, which is pre-trained, so you don't have to do anything. Uh, of course, you're running the mo, uh, the model locally, so you need to have a GPU.

especially if you want to run Cuda here. Um, and then you provide a prompt, and you get the images result. This is what we're going to do here in this case. If you have the necessary hardware to run stable diffusion 3, do it. I'm going to run this one here. Um, all right, so we're going to open up a new python script, and we're going to have to do a couple of installations first. First of all, we're going to do, let me actually see what the Discord package is called. I think Discord py. Oh yeah, discord.py. So we're going to do pip three install diffusers, Discord py, and torch, which is pi torch in my case. All of these are already installed, so I'm not going to run this now, but these are the packages that we need. And then we're going to do some imports. First of all, we're going to import OS, and I'm immediately going to say uh that I want to set the following environment variable TF CPP Min log level. I want to

set this to three. You don't need to do that. I just do it because I get a bunch of warning messages that I don't want to have here. For this video, this is optional. You don't need to do that. Then I want to import uu ID so that I can export unique file names. I want to import torch. I want to import actually from diffusers. I want to import the stable diffusion Pipeline. And I also want to import the Discord package. And what I'm going to do now is I'm going to load the model, or I'm going to specify the model from stable diffusion, or from hugging face. I'm going to say the model, or actually maybe we can copy-paste this from here. So I'm going to say model and pipe. I'm going to copy all of this. I'm going to copy-paste it here. So we have the model ID. We get the pipeline from pre-trained, which is just getting the model from hugging phase. I say I want to run this on

GPU, and then what I do is I create my Discord po. So, I say the token is equal to opening the file token in reading mode, reading the content that is my token, which I got from the Discord developer page. And then, I'm going to create intents. So, the intents is more like a scope of permission or what do I want to do with this bot. So, I'm going to say intents is equal to discord.intents.default. And then, I'm going to say intents.messages equals true, and intents.intents.message_content is also going to be true. And for this line, I also need to have the Privileges enabled, which I showed you. So, this is why we need to enable this here. Um, all right, so we have these permissions, and then I'm going to create a client. I'm going to say client is equal to Discord do client intents is going to be equal to intents. And then, we're going to Define the functionality, and that's basically going to be all we do. We're

going to define two functions. One of them is even optional. The first one is going to be a client event. It's going to be an asynchronous function on connect. This is just optional for me. I want to know that when I connect, I want to print a message bot connected. You don't have to do that if you don't want to, and all the work will basically be done in the second function. So I'm going to say here as well, client event async def on message. So when a message is received with a parameter message here, I can react to it. So I can say, Okay, depending on the message, do something. The first thing I want to do here is I want to check if the message author is myself. So if the message author is equal to client user, so to the user of this client, then I want to ignore the message. So I want to just return because I don't want to go into an endless loop where I react to my own messages. So I want to only react to messages that are not

coming from the bot itself. Um, and then I want to say if the message content starts with, and now you can define a command. I'm going to use exclamation mark gen and space. You can also go with dollar generate or any message. You don't even have to check for anything. I'm going to go with uh exclamation mark gen. If that's the case, I want to do the following. Await and um, I'm going to say message do Channel get the channel that the message was sent in. So it can be a direct message. It can be a channel uh that the bot has access to. And in this channel send the following message. Generating image for prompt. Now actually let's do double quotations here and then message content. But we want to uh to to cut off the first five characters because they are just uh exclamation mark gen and space. So we're going to go five colon um and then maybe we want to use three dots. There you go um. So that is just a message. Now to actually generate the image we use the code here from hugging.

face so we just do exactly this. Uh, so what we do is we say file name is going to be equal, and I do this in order to be able to generate multiple images. Uh, at the same time, or actually, I'm not sure. No, this actually doesn't make sense. But we want to do it like this. So we want to do. Uh, uu ID. uu ID 4.png. Just so we don't have overlapping file names in case someone. Uh, start a second, second request here. Um, and we want to do image is equal to pipe. And the pipe is defined, remember, up here. Uh, pipe. And we feed into the pipe the message content, but only after index 5. So to cut off the, the command itself. Um, and as a result, we get images. We want to get the first one here. And then we want to do image save to that file name, which we generated up here. Now in order to send a message, we want to do with open file name in reading bytes mode as F. We want to say that the picture is equal to Discord file. So we

create a file based on that file stream, and we send that file as a message. So, we say a wait message channel send, and we send it as a file. File equals picture. Uh, all right. So, what's the problem here? No, this should actually work. Uh, all right. And in the end, we want to remove the file from disk. So, from our server where the whole thing is happening. So, we want to remove file name. And in the end, we want to just do client run with a token. And the token is the one stored in the file. So, this already is the whole program. Let me run this and see if that works. So, I run this, and we should get bot connected here in a second, hopefully. Uh, actually, it has to download some stuff first. And there you go, bot connected. So, I can go to Discord now, and I can probably see somewhere here that, uh, what was it called? Image generation bot is active. So, I can actually open up a conversation here, and I can say hello. I should not get a response for this one.

but if I say gen, let's say Doc. I should get generating image for prompt Doc, and there you go, I get a doc. And the same thing should now work also on the server, so let me choose a channel that is, uh, let's actually make a new channel here, text channel experiment. I'm going to do a private channel here, and I'm only going to allow for the image generation bot to be in it. Now I'm not sure which one is, uh, which, so I'm going to do both, create channel. So here now I only have access, and the admins have access, and here I'm going to say gen cat, for example. And since the bot has access to this channel, and it would also have access to all the other channels, so if by accident someone of my, um, of my community members here types exclamation mark gen something, it's going to be generated in any of these channels because it doesn't matter if I send it or someone else sends it. But of course, I'm going to turn off the bot here in a second, so let's delete the

channel, but this is basically how you do that. This is how you can easily build a Discord image generation bot in Python. So, that's it for today's video. I hope you enjoyed it, and I hope you learned something. If so, let me know by hitting the like button and leaving a comment in the comment section down below. And, of course, don't forget to subscribe to this channel and hit the notification bell to not miss a single future video for free. Other than that, thank you very much for watching. See you in the next video, and bye.

Discord Image Generation AI Bot in Python

NeuralNine is an educational brand focusing on programming, machine learning and computer science in general! Let's develop brains!

◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾
📚 Programming Books & Merch 📚
💻 The Algorithm Bible Book: https://www.neuralnine.com/books/
🐍 The Python Bible Book: https://www.neuralnine.com/books/
👕 Programming Merch: https://www.neuralnine.com/shop

🌐 Social Media & Contact 🌐 
📱 Website: https://www.neuralnine.com/
📷 Instagram: https://www.instagram.com/neuralnine
🐦 Twitter: https://twitter.com/neuralnine
🤵 LinkedIn: https://www.linkedin.com/company/neuralnine/
📁 GitHub: https://github.com/NeuralNine 

🖥️ My Coding Setup 🖥️
⌨️ Keyboard: http://hyperurl.co/neuralkeyboard
🖱️ Mouse: http://hyperurl.co/neuralmouse
🖥️ Monitor: http://hyperurl.co/neuralmonitor
🎙️ Microphone: http://hyperurl.co/neuralmicrophone
✏️ Drawing Tablet: http://hyperurl.co/neuraldraw

🎵 Outro Music From: https://www.bensound.com/