Automated PR Code Reviews with AI

Key Takeaways

Hello everyone. Hello guys. I hope everybody is able to hear us. Still people are joining. Can you give me a quick yes? Hi everyone. Yes. Okay. We'll wait for some time so that everybody is able to probably hear us. Yeah. Okay. So Udemy sir Udemy coupon not working. Use August 03, okay? So you'll be able to see the Udemy coupon. So how was the previous session guys everyone? I hope everybody is attending the session continuously, right? >> Yep. Okay. Okay. So, I think just hit like everyone so that we start our session. Uh we'll talk about the agenda today, what we are going to cover and all, right? So, as you all know that specifically auto automated PR code reviews with AI. This is one of the most important thing that every developers will be specifically doing, right? Uh with the help of N8N, we will try to show you that how you can do this. Okay? So, from LinkedIn also, I think everybody is live. I'll let me just go ahead and check it out quickly. Yeah, we're good enough. Like, people are able to see it as great. Uh I was not able to attend the previous session. Do we have the video? Yes, in the live sessions, you'll be able to see all the videos itself. Okay? Great. Hi Chris, the rag bootcamp Udemy are very very awesome. Looking for end-to-end projects. Today only the end-to-end projects has been uploaded already. You can go ahead and see this. PR code means pull request. No, so we'll we'll talk about it as we go ahead, you know? You'll be able to see each and everything. Okay. Now, before I proceed, I really want to make a quick announcement. Uh we have a new upcoming batch, which I think none of you should miss. So, this is the the details. Let me quickly show you. Uh this is the batch that we are coming up with a generative AI for everyone, professionals and leaders, okay? And uh the main aim of this specific batch is to talk about AI, how you can use AI for every professionals out there, everyone in short, whether you are a student, freshers, you are leader working in a specific company, you can definitely go ahead and check out this specific course. Now, this course is actually starting from September 14th, 2025. And this is 13th, but uh I'll update this date to 14th. And sessions will be on Saturday and Sunday, 8:00 p.m. to 10:00 p.m. Along with this, uh you know, there are a lot of things, like who will be the instructor? Myself and Mayank will be the instructor. Who is this course for? Sales, marketing, HR leaders, consultant, and entrepreneurs, operation managers, all leader decision makers, product manager, technical professionals, everyone in short. Here, uh what we are specifically going to master, right? All these specific tools, you know? Here, we are not going to focus more on coding, but in understanding how AI can be used in developing applications. That is what we are going to go ahead and cover up. If you go ahead and click on enroll now, you'll be able to even see the detailed syllabus. So, let me just go ahead and click on this. And once you click on this, >> Uh hello, sir. I think you're not audible. Can you check once? Hello. Sir, I think uh you clicked the button on the mic. Mistake. >> Just now, huh? It is or >> No, no, it's No, it's No, it's audible. So, two two minutes last we I I last we lost you, like the voice only. >> Okay. Okay, perfect. So, now guys, can you give me a quick confirmation if you're able to hear me out? Yes. Let me quickly check. Still not able to hear? Just give me a quick heads-up. Yeah? >> Audible now, audible now. Yeah. >> Okay. So, perfect uh so, here you can actually see, guys, uh we are launching this amazing batch that is called as Generative AI for everyone, professionals and leaders. Uh this batch is all about how you can use Generative AI in your projects, in your applications, many many things as such. You need not even have a coding background. That is our aim over here. So, for everyone, we are trying to make them understand. There is no prerequisite as such. Here, what we are focusing on all is that once you click on enroll now, right? So, once you click on enroll now, you will be able to see this entire page. From this page, you will be able to download the detailed syllabus, okay? And if you have any questions, you can contact our counseling team. From this syllabus, you will be able to see that you have all the information out here. It is a two to 2.5 months program designed for across domain. Here, what we are focusing on? Here, we are easily focusing on no-code AI tools that professionals and leaders can immediately apply to enhance productivity, decision-making, workflow automation. You will master platform for content creation, data analysis, research, and process. Each and everything we are going to cover with respect to the tools. The main aim is that how by using no-code platform we are going to automate all the specific things, right? So, we'll learn about Generative AI basics, prompt engineering, no-code automation, AI ethics and governance, data analysis with EI AI workflow optimization, AI security basics, and change management, right? So, all these things is basically getting covered up. You will be able to see that we will cover generative AI fundamentals, AI tools mastery, professional AI tools, no-code automation, advanced techniques, AI workflow development, and all, right? Along with this, we have a lot of projects that we have planned, like how we are taking this specific project. This is just a basic project itself, right? Similar this kind of projects we will be implementing as we go ahead. Okay? So, as said, all the information is already put up in the pinned comment. I hope everybody is able to see the pinned comment or not. Please do let me know. So, if I go ahead and see over here, you should be able to see I've put the link over here, and I will also make it pinned, yeah. So, it is pinned, right? So, if you go ahead and check it out, you can check it out over here. And you can even call our counseling team in case of any queries, okay? So, this was the quick announcement. Now, it's time problems, any kind of questions, you can definitely ask us while we are teaching the live session, okay? So, everything is available. You can just go ahead and check it out. I've given the page link over there in the pinned comment itself. So, now, Mayank, I think we can go ahead and make them understand what is the today's session. >> Sure. So, sure. So, good everyone. I think we can start I'll just start sharing my screen. Let me know if you face any issues or you're able to see the screen properly. So, let me just add the screen. Great. So, yeah, everyone, I hope okay. Now, it should be visible, right? I hope this same is visible now. So, this was the course I was talking about. I will share the link again as well. So, that all of you can easily access it, right? Or you can reach out to our counseling team as well. The idea is that only. And as I said, the major idea with which we are coming for this course is that even if you're someone that hey, you don't know coding or hey, I don't know coding rag or MCP, but I want to use AI. Okay? So, much like how we use things around us without having a in-depth knowledge, but of course, the same is still helpful. And instead of going around and taking this particular path, the whole aim is that we are able to teach you via these different different tools, right? That how exactly can you master these uh overall things, right? So, that is the whole idea, everyone. And yes, you can surely reach out understanding and what's happening without coding or technical jargons. So, we will be telling you how you can uh use MCP, use rag, use AI into different different everyday tools, and we will tell AI and help you to automate all your tools. Okay? So, you will see that how by the end of this course, you will be saving a lots of time because we will make sure that many things which you're using, be it Telegram, Slack, WhatsApp, Google Sheets, Google Drive, then GitHub, right? All these things we will help you to easily easily just make sure that you can automate them. Okay? And for sure, if you easily just follow the course every week, you will be able to save 15 hours for sure. Because you will be automating your email, your most of the work. Anything which requires just your little bit oversight and see that how AI is kind of getting entered into each and every of these things, right? So, yes, Shahul, we will be using Tableau as well. So, yeah, that will be used for a research-based agent. Okay? And of course, the basics will also be explained. Okay? So, with that, everyone, now let's move forward. And today, what we are going to do, the overall aim for today is that we will be now creating a GitHub PR uh automated automation agent, right? So, pretty sure that most of you will be aware about GitHub. So, this is the GitHub platform, right? In which we can have different different projects. So, I can show you, let's say, one of my projects which we have worked on. Right? So, this is again just some project which I've created for the purpose of demo. Now, whenever someone adds a change, right? So, in a company, what happen is that whenever there is a change which is getting added into the code base. So, let's say we want to make it a little bit better in terms of logic. We want to have some new endpoint added, right? Then, you can just easily add So, this is just a fast API example. Now, you can add your endpoint. You can make sure that you are easily able to use the same. So, the idea becomes that how can I, right? How can I then automate any issues or anything is there. So, just I would like to show you a quick example. Let's say I create a new repository here with change, okay? And in this, I just go and I go to main. And I just change it, right? So, let me edit the same. I'm going to do it directly everything on GitHub only so that it is very easy for you to follow. Now, let me copy this and let me paste it. And let's say I make a new overall endpoint available which is V1. And I am also doing some issues here. So, I'm making some changes. Return uh yes, right? And let me just commit these changes, okay? So, update changes. I can just say new endpoint endpoint added. Right? So, let me commit it. Now, in a company setting, whenever multiple people are working, so of course there will be different different users which are working, right? So, on a particular code base, there can be different different people who are working and can make the changes. So, let's say it's a team of four where you are a manager or you are someone who is in responsibility of handling that how exactly the changes and everything is happening. So, in that setting what we want to do everyone this change which happens. So, let me get the GitHub icon as well. I think it was available. Or some code base icon. Okay, no worries. So, I can just get Yeah, this is the one. Whenever someone get this let's say our code base. So, people will have your code base. And on top of that code base one can make a change. Now, whenever someone makes a change just like how we have updates in our application and everything, we can or someone will have to review the change so that it doesn't fail anything, right? It doesn't fail your whole code base. And for that only what we do is I can now go back to my repo. And I can show you that we can just compare and create a pull request. Okay, so let me just create this pull request which is what is known as PR. Okay. Now, this PR as you can see it has some changes, right? I can see that okay, these are the changes. Now, typically what used to happen everyone is that I as a developer I will come and I will see that okay, what are the changes? Are they good or bad? So, for example, I will be telling that hey, maybe you could have used some other name for this messages V1 and messages is a little bit uh confusing. Right? So, we already have this messages. Then why is this yes here? I think this can be avoided, right? Then we could have added some documentation here as well, right? So, those are the different different kind of changes which we can have here. Okay? So, with respect to that what happened either you will be spending your time in making short of this two and fro between So, let's say this is the user one who got your code base, right? So, I will have this user. He got your code base and he made the change. Now you being the manager, let's say you being a manager, let me just copy this from here and you are the manager here or let's say tech lead. You will be telling that hey, please do this change. Please make sure that you handle this so that this doesn't break. In that way, what we are trying to do, we are trying to now make the manager work easy or overall the team work easy all together. Like hey, all these changes they can just get a pre-vet by AI where it will be able to tell that what exactly are the things which you need to handle. Okay? So AI only will be telling the basic level of changes. It will be checking the logic. Most of the I would say 80% of the things AI can directly help you to do here. And you will be pretty good to easily just then make sure that you do those changes and easy just then move forward and make make sure that you are focusing on the major changes there. Now the same is very very helpful like we use the same into our startup as well where it has been integrated and for the first year, second year or at least the intern it helps me to save a lot of time. Earlier, I have to go and see each and every file. I have to add my comments. For the basics of the changes as well, I have to do those things. Now the same all has been automated, right? And a very very good product on the same thing. Not sure how many of you were still thought of it. So there's a code rabbit right, which is a pretty good similar kind of a thing, AI code reviews. So they have created a whole thing here, right? So if you will see what code rabbit does is exactly the same thing, right? It kind of a PR is created. It adds that hey, what all is the difference here? So they are selling it, right? It's a pretty good thing. They are selling the same as you can see for prices. And it does the similar kind of a thing. Of course with a lots of training, lots of better approach, it is able to do those things. And it will be adding the similar kind of comments. So this same thing we are trying to now make sure that we can do into our workflow. Okay? So that's the whole plan and this is the problem which we are solving everyone. That whatever changes happen and I can maybe just have a another icon. Right? So that you can understand it pretty nicely. So let's say whatever code I am having here, whatever code I am having. Now whenever this code changes, so if this is V1 then I can just have another of this file which will be V2. I say that this is my PR getting generated. So in the technical sense it creates a PR. Right? In the technical sense it will be creating a PR. And I have to make sure that I can just handle all the changes, everything into this PR. So let's see how we will be automating the same with the power of AI. Now doing this with without any tools or with code is also possible, but that is a very very difficult process, right? And this is how we are going to uh understand the problem is now clear. Now we will see that how easily we can just integrate and use the same. Okay? So yeah, that was the problem everyone. Anyone has any issues, anything which is not clear, let me know in the chat. And one more thing, so I was just trying out one very very new thing. How many of you are aware about Google uh this thing, Google Nano Banana? I'm working on the same as you can see. So it has created few of my images from a single image. This is me in 1990s. This is me in 2000s, I guess. This is me in 1970s. So this is pretty funny, though. Right? And this is me with uh okay, let me show you. So let me know if you would like a video here as well. And as you can see this is me okay, let it load. So, meanwhile you can tell me if any issues are there in the understanding. And I will show you that, okay. See, this is me with Elon Musk and Jeff Bezos. So, this is me with Jeff Bezos on Amazon. And this is even me with Jeff Bezos and Elon Musk. Okay? So, I hope you like this as well. We will shortly be going and seeing this. So, we'll be uploading the video for the same in on my channel for upcoming days. And I will show you can use this into your application. So, this can be a very, very big business opportunity, right? I could just easily now say, uh "Can you make the middle guy wear a leather jacket?" And let's see. Okay? So, along with how you can use it here, I will be showing that how can you use the in the code. But in the meanwhile, do tell me if this use case or the problem statement is clear everyone? Then we can move forward with this, right? So, hopefully it will help me wear a jacket. So, that I also look a little cool with Jeff Bezos and Elon Musk. And yeah, anyone has any problem now in the in this particular use case which we have discussed? So, let it generate meanwhile. See? Great, right? Cool. So, video will be dropping on my channel soon, I think by tomorrow. Let's get back and see the same. So, the idea as I said will be that we are able to use it directly. We are able to create a workflow in which we are going to automate this. Okay? Tell me quickly everyone if there is any issues, anything is just not clear and I shall help you with that. Any other doubt as well we can discuss because please stick with the topic so Def Coder, I'm asking the doubts now meanwhile I'm showing this. I'm on the topic only. You see this? This is the topic, right? You should know about the AI updates. Do we have a job out there for no-code tools that we will learn in this course? Uh Rajnish, uh many companies are now kind of reaching out. They are looking for people who can help you automate. We'll just give you an example. So, I was recently reached out with a company and they wanted that these tools can help them just switch from their previous automation tool which are in-house because they are having a good support of AI. So, again, if you think that okay, just by learning a tool you will be able to get that might be not the case, but yes, if you are in let's say non-tech and you learn these tools, your use case will be better and you will be able to do things better. And yes, there's an opening for people who understand these things, right? Same for tech as well. So, companies are looking forward and majorly in startups etc. If you can create and explain these workflows easily, create with a common use case, then sure the same will be very helpful. Okay? Clear, everyone? Great. Uh I cannot automate the CAPTCHA to bypass CAPTCHA for scraping data. Is AI help? Uh Def Coder, I think you will have to explain this a little bit uh like if you want to just uh get through the CAPTCHA, if it's a single uh let's say image-related CAPTCHA, you can use AI. Otherwise, it might be difficult because now those CAPTCHAs are also getting a lot smarter because again, that's the case with AI, right? For any coupon-related thing, anything which is not clear, please reach out to the number so that shall be helpful. Okay? Great. So, since uh everything is clear and we have understood the problem statement in depth, let's now see how we will be doing it. So, for the same again, we are going to use a no code tool. Now again there are many many no code tools in the market like any 10 make relay Microsoft power automate Zapier the one we are going to use is any 10 and I think that all of you have been following the previous videos as well and know why this is very powerful and very useful right. So for this let's go to any 10. Now you could easily just go and click on this get started. I will share the link with all of you so you create the account. The workflow which we will be using here right everyone is already been uploaded so I think the team will share that from where you can download the workflow. So to make sure that each and every one of you are able to follow the same easily we have made sure that we already upload the workflow we created. So this is going to be our workflow. Let me show that one minute. So I can just create a new workflow. And I will show you that what exactly we are going to do there. Okay. So one minute. Yeah. Okay. Yeah I think this is going to be the workflow everyone which we are going to use today. Right and I will walk you through each and every of these steps. So since you will have the workflow already I think the same will be very very easy to follow. Anything meanwhile? I will just share that from where you can download this particular workflow. So if you will go and let's say just enroll onto the free course there we have uploaded. So as there is a single place where all these different different workflows are present. Okay. So I will just show you quickly where exactly the same is. Okay, 1 minute. I think I copied a different link. Okay. 1 minute everyone. So, just let me get this link so I can share with all of you. Yeah. So, this is the link everyone which I'm sharing. If you will enroll for free on this course, you will see the previous videos and everything as well. Okay. So, sharing this with all of you. Great. So, now let's move forward. Now, this particular workflow actually it is a part of an already like it is actually in the templates of any 10, right? So, if I will go on GitHub and let's say if I search here, you will find the template but I will tell you what exactly is the problem there. So, if we go on GitHub and let's say we choose PR review. So, this is again an awesome template like the same template which has been given. But, this particular template is having a problem. I will show you that quickly so that you don't fall for this uh issue. I think they it got uh a problem because of the new uh updates. Now, when you use this pre-created template, right? When you use this pre-created template and let's say when you just try to execute this, I will quickly add my credentials etc. So, pay attention here everyone what we have to do. We will be now authenticating, right? So, this any 10 any 10 we will be connecting with our GitHub. Just like how we do it for other different different applications. We have our GitHub account, right? And in this GitHub account, we will be now using n8n. So, there needs to be this connection via which we can easily make sure that n8n has access to each and everything on my GitHub. Now, for the same, it provides a direct connection where we will be using OAuth 2 authorization, right? So, let me just write it like this. And we will be creating a connection like this. For this, it is asking for a OAuth 2 connection. We could have done the same via access token as well. But normally, I think we all know that how OAuth 2 is a little bit easier to do, right? OAuth 2 is a way, way, way easier to do. So, for that, let's just go to OAuth 2. Now, if you will just click on connect my account, make sure that you are already logged into GitHub, right? So, if I will go on to this browser on github.com, then you will see that how my account, the one which I want to connect is already open. So, this is my account, everyone. As you can see my photo here. Right? So, once we have the GitHub logged in, just go to the workflow again, and just click on connect my account. Once you click on connect my account, it will try to open and it is showing that connection is successful. It will just ask you few things, right? So, what I can do, let me go back. Let me go back. Let me show you by creating another credentials. This time, let me click connect my account. And okay, so let me delete. I think I've already connected this thing. So, I can go on apps. So, your repository settings. Applications. So, I can just easily see that how can we reduce this So, integrations, I think the way it is integrated and all of side I have this any day. So, let me remove this. Let me revoke. So, that we can reconnect and I can show you the same from scratch. Okay? Again, this will be now invalid. So, let me just create a credential again. This time when I click OAuth 2 connect my account, it will be the way you will see, right? So, let me click this. And now see, you will see this kind of a window. Any day by any day wants to access your Mayank 953 account. So, my GitHub is github.com/Mayank953. Now, when I click on authorize any day integration, the same will just happen. Okay, so it's that simple. And I can close this. Now, the same has been authorized. Now, once we have authorized this everyone, I can just select the repository on an owner. So, here you have to specify the name exactly, right? So, if I will write Mayank 953, then it will be able to pick that up. So, it is just a open search box. It is not going to get it for you. And I can just click on this button now to see that yes, my account is getting opened. Okay? So, in that way you will be adding your username. So, github.com/whatever your username is, you will be adding that username so that you can easily connect. Okay? Now, once that is done, then next comes the repository. Now, I am showing you all of these things with this any day demo PR, any day demo PR. So, let's do one thing. Let me just go here again. Let me close this. Go here and search for any day demo PR. And now when I try to execute this, you will see that some error is coming. So, let me first resolve the outstanding issues. So, let me disconnect this. I think this has been disconnected. Right? Let me open this again. Now, when I try to execute this, see, it is showing you an error. Cannot read properties of undefined heating HTTP code. Okay? So, what I wanted to show you was this only that this particular uh workflow or template which is present that is the template which we are also trying to use, but the same doesn't work. And the reason that it doesn't work, you can see right on your screen that how it gives you this error. Okay? It gives you this particular error. So, for the same, I want that you download the workflow by uh which I have provided to you so that you don't face this error when you directly try to launch from the overall uh let's say this uh this template. Okay? Now, if I show you the one which I have given, so let me connect my account again. I think it will be able to connect easily. Yeah. So, now let me go back. And this time when we run this, let's see. Problem. Okay, let me resolve the outstanding issues. I think I will have to disconnect this. So, let's now open it. Let me see. So, okay. I think it here as well we are having this issue. Let me quickly check everyone what exactly we have to fix here. Right? So, I will just make sure that I can get the one which you all have shared because that is totally working without any problem. So, this is the one, I guess. Right? And what I can do here, I can just do one quick thing. Delete this one and go here. Search for GitHub and add a trigger on a pull request. So, this trigger is the one which we have to add. And I can load it here. Or delete for the timing. Now, let me do one thing. Let me This is connected. Let me again add Mayank953. So, this is the one and NA10 demo PR. And now when we execute Okay, this I think has been saved now. Let me execute it again. Cannot read properties. Okay, let us fix this thing. Meanwhile, we see this everyone. I hope all of you are getting this error. Let's see that if you have any doubts, let me tell me please if you have any doubts whatsoever. So, workflow file Nimavat will be there. It will be I think uploaded with a Git with this link, right? With the overall drive link. So, drive.google link it will be uploaded there. You can just check it quickly. Tell me are you able to come to this point and face this issue? What exactly is happening? Because I think everything is given to you already. How to go with enterprise level as code base is shared? So, Jay again, if you are having the enterprise code base, you will just need to connect that to any 10. Okay, but your enterprise may have Okay, it may have some way some let's say things where you cannot connect to it. Right? You Souvik and Jay, what happened is that it has an existing private repositories as well. So, nothing will be there on the open. Okay, so that should not be the case. In that way, you can it is all safe because you can just have a private repository. So, should not be out. Okay, that should not be a problem. So, that way it is safe. That should not be a problem. Tell me everyone, are you able to follow easily? Are you getting this error? The workflow is there. I have just checked in the setup as well. Let me quickly just add it. I see the workflow resource where it is there. Let me just add it again so that you can use it, right? Cool everyone, just check it again. I think I have added the same. Shall be fine now. Just check and let me know. And also let me know if you are facing any issue. I want that okay by this live you all are also doing the same with me. Like I don't want to just do it like myself. I already have done this. But Anypoint should be a third-party tool. Well, JAF or that what you can do is you can self-host Anypoint. So Anypoint can be self-hosted inside your thing. So Anypoint self-host. Right? If you host it inside your company's let's say VPS or anything, then I think it will be all good, right? So easily that can be done. Cool everyone? I hope that all is good now. Tell me in the chat quickly. All of you are able to as I said, the live we are doing so that all of you can follow right. >> Do you have that workflow link the materials people are asking for it? >> What's up? >> The materials people are asking for it. >> That has been already added in the dashboard via drive link. So I think each and every one of them should be able to access it. >> No, no, no. Actually you just send me that link once. I will upload it once. One minute. >> Sure sir. I've shared the link. >> Okay. >> So we have done this. Let me know if you are facing any issues. So that we can move forward with this. And yes, for people who are asking we can easily host self host anything like on to our own architecture infrastructure sorry. So that should not be a problem. Okay. Let me know if you are able to get the workflow I want everyone to follow alongside with me and in few seconds you will see that how nicely we are able to do all these things. Meanwhile if anyone has any question they can ask the same regarding course regarding AI in general regarding anything if you want to have my office views or anything. We will be happy to clear that. Got the workflow. So I think yeah are able to get the workflow. Importance of GenAI nowadays, I think AFK a lot. The reason is that no matter in which field are you in, GenAI is a kind of a skill which is sought after. Plus, if you know how to use it. Now, I know that many people might be thinking of these no-code tools and everything. Maybe as like, okay, they're not useful and everything. But, believe me, I personally have seen that if you use the same, you are able to save a lot of time. Plus, you can generally automate a lot of things, right? So, that is the case. And yes, in that note, it is very, very useful. Great, everyone. So, I think we all can move forward now. Once you click on this, you see that how we are getting this error, right? Now, this is actually one of the error which have been introduced again in a similar manner that someone must have put up a change, right? So, I'm just searching it on Google and someone must have put up a change into the any kind of code base, just like how we are discussing about a PR-related change, right? And actually, this error got induced. Okay, so you can see GitHub trigger cannot read properties of undefined reading HTTP code. So, if you're a forum kind of a guy who can read through this the way I read through this and was able to solve this error, you will be good. But, if not, then of course, I am the person who can help you with this, right? So, there again, they are saying that hey, some issue is there. So, how to solve it everything. Now, they have created a pull request for this. If you will see, they have created a GitHub issue for this, where also they are having a pull request. So, you can see the pull request here. And the same thing which we are kind of trying to do, error HTTP code, they did these changes. And I think in that, something changed. Okay, so that is the whole problem. And yes, that is the reason which by which we are also focusing on how can we use the same properly. So, now what we will do everyone pay attention. So, let me close this. Let's go back, right? Let me just leave this. Let's go back. And I'm going to save this workflow. So, you will be easily able to get this particular workflow via the JSON which we have uploaded. Now, I'm going to create a new workflow. So, pay attention. I'm going to create a new workflow. Now, once you create this new workflow here, do GitHub. And in this search for pull request. So, on pull request. Now, we will try to do the authentication again. So, just do the authentication again on this new sheet. Just make sure that you properly do this. So, this has been authenticated now. Now, what I will do pay attention. So, now I will again select myself. So, Mayank 953. Right? Mayank 953 N8N. So, what was this exactly? N8N demo PR. And now when I click this, let's try. See, this time it is listening, right? It is not giving me that error. It is saying that hey, I am listening for test event. The second you create a PR, now that error has been kind of handled where it will be now checking that. Okay, so I was able to understand the same by going through this particular PR and everything. But for our use case, I think we are pretty good. So, what I can do I can just go back and I will just quickly create a new PR. So, just uh pay attention, okay? I would just like to create a new pull request. Let's say I create a new pull request. That pull request will be base will be main. And I can maybe let's say select the PR example. Okay? Let me now create a PR PR. I think okay, this has already been created. Let me try to close this pull request. I just want to show you by creating one that how exactly our N8N will be able to catch that PR. Okay, so I will select PR example. I will create a pull request. Now, once I create this pull request, it will send a signal to my N8N that, "Hey, something happened." And what was that something? That something was this thing, that it is able to now easily get the details. Okay, so many, many details it is able to get. And this is what we want to use. Okay? So, that is exactly how why we need to make sure that we are able to solve it. And the way to solve it, right? The way to use GitHub trigger in N8N till they solve it, is in this manner only. Now, what I can do, I can just go back, right? I can just open the full workflow again, the one which I've given you. And I will be now just copying the other notes. So, let me create another new one. Let me create new one. New workflow. Now, this workflow gets created. I will just open so the file which you are getting, you can import from file, or you can just copy paste that file here, like this. Okay? Now, once I've copy paste copy and got the file, what I will do, I will just delete this particular thing. So, I will delete this faulty node in the workflow. And I will just copy all of them. So, let me copy them, go back, and the best part is that you can genuinely just copy paste it like this. So, see, the whole thing is now together. Right? I'm able to get each and everything. I will just move them and connect it with my this particular flow, thus making the overall Now, once we able to fix get file difference from PR. So, when we have the file difference which need to be worked upon, and create target prompts for PR differences. So, here we are just setting the prompt, okay? That what exactly do we want it to have? So, for example, here again some code is there. You can just move forward. And we can see, you are a senior Python developer. Please review the following code changes in the files. We will get the differences. Once we get the differences, then we can just discuss on the missions, review the proposed code changes by file and by significant modification, generate inline comments, ignore files without patches, do not repeat the code snippet or the file name. Write the comment directly without introducing the context, okay? So, with that, now once we have done this, then we will move forward with this agent. So, this is an AI agent everyone, which we will be using. And it will be having a brain of its own. So, in this, we can easily just, let's say, use and claim this free credits. So, when you create a new Anyscale account, you can just claim these credits. Or if you want, you could just maybe use any other model, totally on you. You can use a Gemini model, you can use anything, totally on you. Okay? So, I'll just claim these free credits. And I am selecting GPT-4o from this. Uh so, let me get GPT-4o mini now. Okay? So, these are the models which are available. Then, we will ask or since we are connected with GitHub, we will try to add back whatever changes the AI is suggesting, the agent is suggesting, back to our overall PR. Okay? So, that is going to be the flow now. So, everyone, tell me now, are you able to follow this properly? I think this is a valid use case each and every one of you shall be able to do the same, right? Let me know if you have any doubts, if anything is not clear. Grok is also supported in Coder. So, you just, let's say, I delete this particular uh connection. Go here. Select Grok. And we can get the Grok chat model. So, see, that easy. Again, for the credential, go and create a new credential, add your key. And yes, I think we all are good here. I will just check if I have the key here. So let me just check with my key. Save. And seems like the same is working. So now since this is working everyone, I could just select any of the models totally for free as well, right? So the Grok, Open Router, all of these are supported. So you don't need to have that. But I think the best part is that how Anytown also gives you these free credits, right? So I will stay with OpenAI. Great. Any doubt anyone? Let me know quickly in the chat if anything is not clear. I will help you to quickly clear the same. Again, it's a very, very helpful use case. You can surely just implement the same in your company and believe me, your company is going to be very happy. If I would have had this, uh lots of time 2 years back must have been saved in my life. Okay? Let me close this and just zoom in. Great. So now since the whole workflow is clear, okay, what we will be doing everyone, we will now just try to understand and just execute these workflows one by one. So this is a API which GitHub provides where we can get via the PR that okay, what exactly is the difference between the files. So let me just click this. And in the table or let's say via schema only, we can understand that okay, this is an awesome repository regarding Anytown tutorial for GitHub. Now if I go back to my GitHub, and if I go back to the file changes, you can see that this is something which I have added in my readme, right? So, readme plus one. Now, if I again go back to my workflow, right? This has got all the files. Now, next is create target prompt. So, this code will just try to understand. Uh for the timing, if you don't understand this code, should not be a problem. Let me just execute this. We have just given it some prompt and just asking that, "Hey, this is the user message. You are a senior Python developer. Please review the following code changes in the files." So, I can show you the payment JSON. This is the file which is readme.md. The difference is this create this is an awesome repository regarding this and your mission is already given, okay? Now, I will go to the next one which is the AI agent. I will just execute this, right? Let's me execute this. And this agent will take up this user message and tell me that, "Okay, what exactly are the changes which it suggests?" So, it say it is saying, "This message is a good start, but it could be improved by adding more details. Consider mentioning what any tail is, how this repository can benefit users, including installation instruction or a brief overview." Now, if you want, you can just give it some more information here that, "Hey, this repository has to do with fast AI, not with MCP." Uh please make sure that you are properly giving the same in points. So, let me just give it you are a helpful assistant. Please give your analysis in points. The repository has to do with fast API. Now, when I execute this, see how the output will change because we have told our agent that, "Hey, something is different." See, the readme file is newly created and provide brief description. Description mentioned this which may not accurately reflect the content if the repository is primarily about fast API. Consider updating the description to better represent the project, okay? Now, since we have done this, let me move forward. And now, it is not able to get this so let me fix this. We can just go back and change this to PR review. The reason was that in the original workflow, so this was PR trigger not PR review. So let me fix this, okay? Let's go back and let me rename it. So I can just press space and rename the same PR trigger. Now see this has become green because it is able to get this particular thing. Now when I execute this, let's see. It is able to get all the information. It is able to get the body as well. If I go into my PR, let me go in here. I think nothing has been done till now. Now when we go back on to our workflow, then the last one is the add label to PR. In here when we click this, as we can see, this has done the same. And when I go back to my PR and let's say refresh this. Uh one minute. I think it was the different PR. Let me check which PR it was. So create readme. No conflict commits. Files changed. This is the file which has been changed. Let's check have we missed out this thing. Where exactly it is supposed to do? So let's go back. Mayank953 in in our demo PR. So where exactly it has missed it out everyone? Let's see the same. Or let me do one thing. Let me just create uh let me just Now since we have understood the workflow, let me execute this workflow. So it is not waiting for the trigger event. Let me go back. Let me go back to my main. Create a new PR change here. So I can I will just create a new PR with changes, okay? In this PR, I will just quickly go and let's say I will just add a readme here only, okay? This is an awesome repo for a n8n changes. Now when I commit these changes, I can commit these changes. And I can go back and create a new PR. So, if I just show you, let's say if I just select 1 minute. So, what I can do I can just select this. Tile vertically. So, let me execute this workflow listens to this. See, it is waiting for the trigger event. Let me compare and pull request. Right? I will just create this PR and pay attention. I can actually make it active from here. So, pay attention. I can just make it active from here. Once I make it active, right? Then it will be listening. So, it will be just listening and if I create a pull request, so the pull request has been created. Now when we go to execution, we will see that how an execution has started. So, this execution is running everyone. And see how quickly it has just added. This line introduces a new readme, which is crucial. Additionally, this line introduces a description of the repository. However, it lacks the detail. And this is exactly what was given out, right? So, if we see here, this line introduces a new readme. It is easily able to now just add it that quickly. So, I hope you are all able to understand. I think it got kicked up two times, but that should be fine. We can check why it happens. But yes, as soon as I create a PR, it is showing me that okay, what exactly happened, right? So, that is the whole benefit. Now to make it a little bit more better, what we can do, pay attention again. I can go to my repository. Let me know if anyone has any doubts now. Let me meanwhile take the doubts. Can you click on the GitHub trigger node again? I can surely do that. If I just go outside, I can go on editor. And in here, this is the GitHub trigger node. Okay, everyone? See, I've just done the authorization. Uh just make sure that you start this up in the new one, okay? Not take that overall uh example. Just that. Will company allow to use uh n8n? If you learn about how to use n8n on a self-hosted infrastructure, which is very very easy via the Docker image, I think right now companies can easily allow that because again, nothing is going outside from that, right? So, if you're and you can genuinely use AI, which is local hosted as well. So, let me know if you want a video on that and I will help you with this, okay? Can someone pursue career in Chennai or in India by only learning this no-code way? What career option might be open for an individual if the course is taken on knowing n8n? So, Rupajit, I will be telling this from an experience of mine. Like it started from a technical perspective. If you only understand this uh no-code thing and let's say you want to go through the technical field, then it might be very difficult. That said, if you want to now using these tools, if you want to now go into the non-technical or a semi-technical field, that might be still doable. So, proper tech, I think it will be a problem because somewhere or the other you might have to go beyond white coding, right? So, let's say if you're in sales, HR, marketing, if you are even, let's say, leading a team after 10-plus years where you don't code, but you need to understand that, okay, what exactly are these different different concepts, then the same can be helpful, not the other way around, okay? Especially in companies like JP Morgan, etc. So, again, Tech Dose, I think I know from where you are coming from. Uh Goldman Sachs, where I have worked, that was also very very particular about what exactly are they using, but I believe that if you can host and show them via their infra and show that why the same is useful, how the data is not leaving onto a locally hosted AI, then I think they should not be a problem, but your presentation, your way should be like your overall confidence should be very high. Like for example, if I have to convince my managers now, I think I could easily do that. Right? I could easily easily do that thing. Because I know that how it can be hosted by using a simple own AI. So, with that knowledge, I think I can now do that. But of course, if you don't have that, then it will be a problem. So, showing that how data is not leaving, everything is locally hosted, how it is very helpful, it can help you do so many things, then it will be better. Because doing the same via code, that can be a problem. Can I create AI tutor which kind of AI tutor Pinky Sahu Pinku Sahu? Great. So, meanwhile this comes. Now, let me just show you the best example of this. Let's say this particular thing is still happening. What we do is like I will be working on this code base. So, let me go to main. And I will be doing the same via some VS code on IntelliJ, but I can create, let's say Mayank changes as my new branch. Now, in this branch, I will go to fast API template and I can go to, let's say main.py. Now, I can edit this. Right? I can just do the same thing which I was saying to you in the initial that, "Hey, now I can add a message one V1." Something like this. I can do like, let's say if I am sure or I don't know the coding, I do a lot of white coding. It can just print something here. Print Hello world. Right? Then I can just commit these changes. Right? I can go back to my this branch only I'm making all these changes. Right? So, let me go to the branch. Let me just go to models.py. Let me make some edits here. And I can again do some class not usable or error here. Right? Error message string. I can just add this as well. Right? Now I have done few changes and in this particular uh PR, I can just now create a new PR. Okay, so let me just create a new PR. So my own changes, let me create a pull request. And let's see the same in action now. See, I've created a pull request. Each and everything is happening. So let it run. Checking for the ability to merge everything. Right? I have two commits, two changes, but see, as soon as I created my PR, right? I have not done anything. I have not changed anything. My workflow is running. The workflow is pretty active in the backend. It is able to tell me that, "Hey, the new route introduces a new PR. Right? This is significant addition to API, but naming of the endpoint could be improved for better restful uh structure and clarity." The function add message now contains logic to generate unique ID. The approach of using this works, but could lead to issues. So the addition of hello world, right? You can see here, the addition of hello world, that also it is able to tell that how it is not very helpful. Right? So let's see the same. The addition of hello world is unnecessary for production code. Should be removed. It can clutter logs. Consider using structured logging if needed. The message payload model now includes a message ID. So I hope you can see that how it is telling each and everything right in front. Right? The error class is not defined as a Pydantic model, which should be the case. So now think that even if a fresher or someone who doesn't know the code, right? How easy it is like normally it is giving you these messages. Right? And this is the benefit. I could have the first level of checks very, very easily done. And this is again what this code habit kind of does in a hindsight that you can easily create this. Then it tells you the issues everything going a lot inside your code, right? But with respect to this particular thing it can easily get so many things for you. Consider adding a response model. URL can be benefit. This line should be removed that is does not fit within the fast API application. So all these things is directly being happening saving me a lot of time. Now of course you can have a more complex use case, more complex changes and everything will be handled. Okay? So yeah, with that we have just completed this. If I go on the execution you can see that how the execution happened, right? How the execution happened it is being completed. If I make it inactive now, it will be fine. So yeah, I think this is where the overall power of our GitHub AI automated PR come into picture. We have also seen that how can we correct the error which we normally get. How the same is useful in real life. As soon as I create a PR the same gets triggered and it can just handle it. So think of this that in a company setting where many freshers join or everything how helpful this is because otherwise I have to go and understand the changes. And now at least a vetted PR can come to me or to the higher management just making the work of tech guy a little lot easier, right? And even from a non-technical perspective this will tell a lot of things, right? So that is the whole idea. Great. So yeah, with that I really hope that everyone we have understood the same and you can easily just get this workflow downloaded again if you want. The same has already been uploaded though so you can use it directly. Just make sure that you are able to fix the error the way I have told you in this particular video. Great everyone. So let me know if you have any doubts anything which is not clear. We'll be more than happy to do help you in that. And I hope you like the project, right? I think the project is really helpful as I've shown you by showing the thing live. So, let me know if anyone has any doubt. Private tutor, I want to create a private tutor that helps you to solve their problems. So, Pinku, what we can do there that's a very good use case. We can create that, right? We can create a private tutor which can talk and converse with the student and can solve the problem as well, okay? Bro, can you create It is possible to get an AI agent which can tell you how much views, likes the content will get? I don't think that will be very valid. It You You can just give all the YouTube data and that will be helpful, okay? Can you make an AI agent which uses Nano Banana? There's an option of adding image of a model for consistent face and option for dress and pose. Yes, I'm already working on that. Hopefully, you will see the same product on my channel tomorrow. So, as I showed you, I have actually after spending lots of time I have worked it out. So, there are again different different ways through which you can use Nano Banana, right? This is the one which normally you would see. So, let me go to Google and from there I can show you. Nano Banana API. So, either you can use it directly or you can use via this file.ai. Now, via this file.ai it is difficult because it asks you for credits, but don't worry I will be creating the workflow. So, I have actually already created one where you can see full workflow with image, without image. I'm just testing out the same. And yeah, then it will be pretty good. So, you can just upload an image and you want You can just tell what you want. And finally, we will be able to see that. So, let me just show you a quick example. Let's say change the logo to Amazon. Make me wear a jacket and add lightning. So, let me submit this. Now, again it will take your image, change the image. All these things will be happening, but let's keep on to the today's use case, right? The today's use case was also a lot helpful. So, now see the same is happening. So, the major challenge was in getting the API and everything, right? That was a major challenge here. And once it gets the file, I'll show you that what it has done, okay? So, let's see the result view. Okay. Not sure. Okay, I I I meant the other kind of lightning, but it has done some other kind of lightning. But, I like this thing very nicely. Let's see how nicely it has integrated the jacket, right? Even if someone showed me this picture, I will also get confused because I have this kind of a jacket. That's another problem. I have this similar kind of a Levi's jacket. So, this is what I like. This Amazon is what I like. This lightning it went a little bit overboard, okay? So, I will be showing you this workflow as well. Don't worry. Cool. Let me see the questions. Can we add another agent on top of this? Yes, uh Raja, we can add any kind of agents. We can at least We can have a multi-agent system. So, that will surely be done. How accurate is it since it's AI? How much can we trust it? So, Shaman, again, uh should be fine. Like, if you just properly prompt the same, it will be very easy and you can trust the same, okay? Helpful session. Thanks a lot, Sridevi. Uh okay, manual intervention will be required, I guess, even though it is automated. So, Shaman, again, if you will see this in action, right? It is not in any case it is doing something. It is just getting the comments. So, if even if the comment is not uh clear to you, you can surely go to and go to your tech lead or tech person or the senior and get that, okay? Lead generator for exporting Indian handicraft products. So, Pushpalata, that can also be done. Uh I will be doing that end to end in either one of our courses. So, the course which we are doing everyone, right? And you can use Mayank as a coupon code. It will give you 15% off. Right? So, for people who are staying till the end, we give this coupon. So, that use Mayank for 15% off on course. Right? So, I think this will be helpful for all of you. Let me mean basically more comments. So, thanks a lot for staying. Uh I will just be taking for five seven more questions and then we will close it, okay? Please ID on AI tutor. So, Pinku, I will be uh creating that and I will upload the same. Don't worry, okay? So, we are slowly slowly moving forward with that. Can you explain what is an AI agent? Uh okay, I've done that multiple times, but I will suggest for people who are not aware about any 10 in any case, what I will highly say is just go uh you can go to my channel or even Krishna's channel as well, where we have uh explained about this AI agent and what all is this. Parallelly, what you can do, I have a master class specifically focused on any 10, right? It's a 7 and 1/2 hour class. You can just see this, right? It has 122k views, 4.5k likes, 180 comments. So, doing pretty well, right? So, you can see this class where I've explained each and everything in depth. So, about automation agents, what's exactly they are. And the use case as well, right? We have created so many different different workflows here, integrated the same with Telegram. So, all that thing has been done here, okay? So, I will share this with all of you so that you can go via this. Which course are you talking about? Uh I think the link has been added in it, so you can go through that. Okay? Are you using a paid version of n8n? Tech those I don't use the paid version of n8n. I have the n8n for my clients and the way I do this consulting, I have the n8n hosted on a VPS. Okay? So, on a virtual machine we have that. Just after you n8n master class video, there's a n8n so that will be really helpful. Uh which course are you talking about that is done? Then good one which no-code AI tool can be used for developing a website from scratch. Again, Krishna, you are getting confused. Website is a different thing. Website you can use maybe Lovable or some drag and drop kind of a tool that is front end. This is for workflows and for the back end majority. So, n8n is not giving you a website, right? So, for that you will have to use something else. So, please your 7-hour video dubbing in Hindi. Dubbing to Pinku I'm not sure, but I can create one in Hindi if you want. Okay, dubbing doesn't give that fun. Is everyone let me know if any doubt is there with respect to this workflow, with respect to anything? First of all, thanks a lot for the person who created this. But since we were facing this error here, so I made sure that I can correct that error as well. Okay? So, now with the error corrected, which was there in the official workflow, I hope all of you can see that how nicely we are able to get things done. Okay? And yeah, I think we will we still are using this and we will integrate this into more of our use cases. Okay, hello, sir. Yeah. Okay, guys. So, do definitely check out the generative AI for everyone course including professional and leaders. Everything is given in the description of this particular video. Uh we are currently running 10 to 20% off, you know, and I think there are only 150 seats that are remaining. Uh the batch will basically start from 14th of September. All the information is given in the description of this particular video. So, yeah, this course can be joined enrolled by non-tech anyone. You don't need to have any coding knowledge because there we are not going to focus much on coding. Okay? Perfect. I think we can stop the session. Uh >> Cool then. Yeah, thanks a lot everyone. Just reach out to me if you have any problem. And I think I will be more than happy. So, I'm just sharing my LinkedIn and YouTube quickly so that you can easily reach out to me. Okay? If anything is there. Yes, sir. I think we can close the session. It >> Yes. >> Before we were able to >> Okay, guys. Thank you. Find out all the information in the description of this particular video. Thank you. Bye-bye. Take care. Thank you.