Build your "OWN AI AGENT" for free | No coding required | @tech.boy.deepak

0:00 So guys welcome back everyone it's been a long time since I made a video and I know that we are all concentrating on this. You will know why I made this in the video and as we said in the thumbnail, even if you don't have any coding language or coding knowledge, you can build your own AI agent whether you are in the first year, second year, third year, fourth year or any group. So I'm giving you a small preview of what this AI agent does. If you are running the agent, enter the prompt like this and hit run agent. Let's wait and see what the agent responds to. And I didn't use any hands. Now I just gave a prompt and searched for who is the founder of instagram. I clicked the same search button and clicked enter. So we got the founder's name,

0:45 copied it and gave it to us in the result. So you see, in this way, we don't use any hands on. With just one prompt, we can work with our AI agent. So, in the future, many more big things are going to come. For that, it automatically books tickets. For that, it automatically orders from Amazon. So, these agents are definitely used by many companies to reduce their workforce. So, for that, watch this video that I have made explaining the hands-on video completely. And before the end, at least 80% of the students should build an agent completely. And if the remaining 20% ​​of the students have any doubts, if you put your doubts in the comments, I will give them clarification. And let's build an agent.

1:31 Let's see how many people will complete it and how many will comment. And do subscribe. Subscribe for such up-skilling content for such videos. Let's dive into the video. So guys. The recording has started and everyone who has started the video that was being informed in advance, every command is very important. You have to build your own AI agent, that is, watch the video until the end and follow the steps that I am following. Many people are concentrating on this. I found this in Kumma. So if we put this in, we will get 100k subscribers. That is why I put this in. So as you can see in the thumbnail, we can build an AI agent without taking any API subscription or coding knowledge. I will show you how to do it in this video. So all

2:19 of you open your laptops and put it in. The first basic requirement is that we have to make an AI agent. The basic requirement is Python. So first download Python and download the latest version. In Python downloads, the same process will be followed even if I am using a Macbook or you are using Windows. Almost but where is the difference, I will also tell you that difference, you follow them, so I am downloading Python, okay, so I know that you already have Python on your laptops, you are coding, so continue, continue, continue, okay, just install it, so yes, Python, Python is started,

3:05 and in this gap, I will show you what we are going to do, so first, the technology that we use is browser usage, so I will also explain everything, the terminology, let's see what browser usage actually means, so browser usage is the easiest way to connect your AI agents with the browser, when we create an AI agent, it is already an open source repository, so we are creating this agent using this repository, so in this video, first, you have browser usage, we have two videos with links, right, GitHub links, first, the browser usage Related to the second, the browser uses the web UI, these two should be opened, open link in new tab, open link in new tab, so the browser uses the GitHub repository, the

3:50 browser uses the web UI, all these links that I have mentioned in the description, you can also accept them from the description, you can access them all, so do you understand, and the browser uses the repository is already available in open source, we can make an AI agent, just clone it and see, so the browser uses the repository, first I am in the quick start, okay, so let's start from here, first we have already installed Python, and then next, pip install, browser uses, okay, I am using this command, okay, Python will be installed, let's wait, and at this time, if you are also installing Python on your laptop, hit that like button and subscribe, because small

4:38 apps like this mean If you build an AI agent, you can add it as a project in your resume. I built the AI ​​agent using LL.Ms. You can take it to an advanced level. So videos like this are very helpful. Everyone has no connection to the group. Anyone in IT, EC, or CSC can build this AI agent on their own. So follow each command exactly. Let's see how many people build the agent before the video ends. Let 's wait and watch. Python will be downloaded. Installation. Then, make sure that the internet is correct. Connect to WiFi. Otherwise, make sure that the mobile hotspot is also properly. About a minute, okay.

5:29 Python installation is complete. Close the Python installer. We don't need Python. Python is installed. Now, throw the installer in the trash. And in the terminal, our first command is pip install browser use. So install browser use. So okay, it says pip command not found, okay, we should use pip 3, we should install pip 3, browser use, so let's see what is the difference between pip and pip 3, pip vs pip 3, what is the difference, pip refers to python 2, pip three refers to python 3, so we have downloaded the latest version of python, so we should install pip 3, so browser use is installed,

6:14 next, let's go to the git hub repository again, next, let's see what is play right, install play right, once play right, okay, so play right is just like this, this is also an open source automation library, now as we saw in the video starting, if we just give a prompt, which agent will complete the same task for it, so play right is an open source library that automates that task, so it is clear here. Play Right is an open source automation library for browser testing and web scraping developed by Microsoft. So when we give it a prompt, it opens Google and asks where the search bar is, what to click, it automates the whole thing. Play Right has to remember all this.

7:01 If you add it in your resume and ask for an interview, you have to explain all this. So, I am also telling you the basic basic things. So let's install Play Right. Install Play Right. Okay, so wait for this too. And we are going to use Gemini. Gemini. Gemini API in this video. So, do you all know what LLM is? A large language model is a machine learning model that can process and analyze natural language. So, we are already using Gemini. Google LLM. So, we just need to

7:46 use the API. We don't need our own LLM. Play Right will be installed. Wait for each command and install it like this. While it is happening, do not do any enter or go back. If you do go back, it will stop and cancel. So just wait. The time will run here and okay. So it will be installed. Play Right. And before that, go to the comments and put the original. This AI agent you are building. So what have you learned? If so, after this video is over, what have you learned? Is this

8:32 video useful? If not, what kind of modifications can we make and how should we explain? Do you understand my way of explaining or anything about it? You can comment. And one more thing, if you have any doubts until the end of the video, put that doubt in the comments. I or my team will respond. No worries. So I think Play Right is installed. No, it is not. It will be done. And then. So we have installed Python so far. Play Right is installed.

9:18 Yes, so Play Right is installed. 638 mb. Tony Done Play Right is installed and now create a folder on the desktop New Folder and let's name this folder also Rename it So this folder name comes AI Agent So I named it AI Agent Just recording is working Done So I created a folder name AI Agent And now we need to come to the desktop First we know our terminal commands If we type cd means change directory desktop Our terminal will come to the desktop And next

10:04 on the desktop we have created a folder The folder name comes AI Agent So cd space AI Agent If you type this you will come to the AI ​​Agent folder So in this AI Agent folder we will have a browser use Git repository So in this browser use we have to clone the web UI So browser use web UI So this is the second Github link that we took from Google So this web UI is here code Clone it and here so we clone the repository What we need to do is use the command git clone paste enter so done clone is done now if we open our folder we will see something called web UI so we cloned it directly from git hub

10:51 and done and next step comes and let's go inside web UI the folder inside cd space comes and let's see what web UI web UI next next we have cloned web UI now all the commands that I am reading are in web UI repository that cd we have entered web UI web UI next step comes and we have to create a virtual environment for our AI agent to run in virtual environment and before that we have to install u v we are here step to we recommend using u v for managing the python environment let's see what u v actually means so u v v means it is an extremely fast python Package and Project Manager is written in Rust.

11:37 This is a Python package. You do n't need to know anything about it. Just install it. So, if you are copying the command, it is here. So, for those who use Mac and Mac and Linux, this command. For those who use Windows, this command. So, since I am a Mac, we are pasting what we copied. Do you need to paste it in the web UI? Remember, you only need to paste it inside the web UI. So, it will download. You will see that everything is installed. Success. So, next, you are done. And then, now let's create a virtual environment. Copy. So, it says that you have created Python 311. So, let's paste this link. So, by

12:23 doing this, we have created a virtual environment using Python. So, we have created a virtual environment. So, we need to activate this virtual environment. To activate it, we have given a command here source dot v env / bin / activate so we can see it here too so if you are using Windows then you should use this command to activate it if you are using Mac then you should use this command so we are activating the virtual environment so how do we check if the virtual environment is active or not i.e. here before we see the web UI like this so when it comes like this the virtual environment is active after the virtual environment is active then let's see the next step here again step three we have to install dependencies we need to install

13:09 what dependencies we need to run any agent to install them so you v pip install hyphen r requirements dot txt ok requirements dot txt it asked to install the file done installing same same Copy the command and paste it in our web UI. Okay, so the requirements.txt has been downloaded. So there are many libraries in this requirements.txt. We have downloaded all of them and then it asked us to install the play right. Again, we have already installed it. Let's do it again. Play right is installed. Done. We already have it and we don't need it with the next docker. We are doing a local setup. So this one local setup is

13:54 here to run this web UI. This command has port 77 and 88. Take this command. Then paste it here. Okay. And after pasting this, if we wait for one minute, an IP will be generated. If we take that IP and paste it in Google, we will see that the interface with that web UI is real. So wait and if you have any doubts till now, put it in the comments. Or watch the video in slow motion. Everything is clear. The step is shown and the document is given in the description, how to follow the steps, what libraries should be installed, let's wait, it will be downloaded, so we have created a port, let's

14:41 see it in the virtual environment, okay, so run in local URL is here, copy this, http / and this is what we have done on our computer, we use it ourselves, others do not know, so the local setup is done, and that link is being pasted on Google, so now what is expected, so we have used web UI, what is web UI, that is, the technology we have used to run that agent, that AI agent, is browser use, but we have to give a prompt, that interface should appear, right? That interface is web UI, so now I just hit enter, we get a user interface, we have used a technology called browser use,

15:26 this is web UI, so we have this user interface to run an AI agent, so here is a lot There are settings, agent settings, configuration, browser settings, run agent, deep research, so first we have to configure the LLM, so the LLM that I am using in this video comes from Google and in it there is Gemini 20 Flash, so how do we get the LLM API, that is, we have to go to Google Studio, Google Studio and then here it says Get API Key, I am already logged in, but if you are not logged in, log in for free, all this is free, we do not need to pay anything for this, create API key, and we are using Gemini API, click on Gemini API, click on Create API and OK, we have an API key, it has been generated,

16:12 this has also been clearly mentioned, do not share them or embed them in the code, the public can view it, so it is not for the public, this is for your API, so I copied this, copied it and entered it in our browser, this is here in the API key, so we are using the LL name, Google model name, Gemini 20/ And the api key came and we took it from google studio and then we don't need these browser settings it's already default run agent this is our main main page from here we give a prompt to the AI ​​agent and it will generate it so let's see go to google dot com and search for who is the

16:57 founder of instagram so let's search like this let's see so if I am running the agent enter the prompt like this and hit run agent let's wait ok I hit G O O LE are we bluffing and watching what the agent responds let's see what G O O LE com is open I think ok googlecom is open so it understood the agent and c I didn't use any hands now just gave a prompt and it searched for who is the founder of instagram and then let's c so it clicked the same search button for that it

17:43 clicked the same enter for that so we got the founder name Let's copy it and see what it gives us in the result. So once your AI agent completes the task, this website will close. So when it closes, you can check. Let's wait. The agent will do the task. Um, I think it's done. S. It's closed. And here on our terminal, we get the message that the task was completed successfully. So once we go to the web browser and check the results. In these results, it also shows us the latest recording of how we did what we did. So here is the final result. Suggestions have appeared below the input field with different variations of the search query. The future snippet states that Kevin

18:29 Systrom and Mike Krieger are the co- founders of Instagram. So this is the result. So what we have done now is that we have created an agent. If we give it a short prompt, it will go to Google and complete its task. And we can do this. How can we take it to the advanced level? In the future, new models will come with new advancements. For example, if you want to book a flight ticket, just give a prompt, it will go to Make My Trip or any platform and book the same flight. If you want to book a hotel, you can go to an agent directly. You can write an email. If you want to write a prompt, it will go and write an email. Everything will be like that. So let's try it once. Does it write an email? Does our model have that capacity? Go to google.com and open mail and

19:20 compose a mail to Deepak [email protected] by informing him to attend the AI ​​event. Open mail and open account tech boy Deepak @gmail.com. So now I am looking to write an email from my mail to another mail. Let's see if the run agent is running. Let's check whether it actually writes or not. Let 's do it, so if Google opens, it does it, okay, next, we should

20:10 open the mail as we thought, okay, Gmail is opened but it asks for sign in, okay, so if we didn't enter the password, I know it won't log in, it tries to log in, but I know it won't log in, let's see what happens, so if C opens Gmail and Next

21:14 searches, and so in the future, with the fine tuning models that we will get, it will be fast, they will not be this slow. For example, if we need to book train tickets, that is, Tatkal tickets are opened only at specific timings, the same agent will do the same for it, C entered my mail, so I click next, now it has no password, let's see what it enters, zero, so I know now it goes back completely, it

22:02 tries, it clicks on Try Another Way, there is no password, so I am canceling, so I stopped there because we did not give it a password, so this is Our own AI agent Even if we don't have any coding knowledge, we can build an AI agent by just understanding these commands and knowing where to use which command. And here till now you have watched this video, I think you are also building your AI agent and after you build it, you give a task, what was the task you gave to the AI ​​agent, did it do it or not? Tell us in the comments. Let's see who has tried it in how many different ways. So that's it. In the future, there will

22:47 definitely be a good scope with these AI agents because many companies are taking away employees and working with AI agents. So if you are an AI agent builder, if you are a developer and are doing it at a good advanced level in making AI agents, you will have a good scope. So I think there is a good scope for this. And if you build it, you understand it and make a request to everyone. After building an AI agent Go to the link and post it. I built an AI agent using this browser. We have done this so that our friends also understand that we have built an AI agent on our own so let's see how many people will build it I think more than a 100 plus

23:32 people but I think they will build an AI agent on their own now let's see let's wait and watch and that's it so we have developed an AI agent for free as shown in the thumbnail and that's it thank you so much for watching keep supporting