Claude Tutorial for Beginners: Become a Pro in One Hour

[00:02] tutorial. My goal today is to take you from zero to confident in terms of using Claude. You don't need to be technical as long as you know how to send emails or browse internet, you should be able to follow along. Claude is an AI tool

[00:16] similar to Chat GPT. It is created by a company called Anthropic and it has been rising in popularity since past many years. I personally use Claude more than Chat GPT nowadays and I'm pretty sure by learning Claude, you will gain lot of

[00:32] benefits in your career as well as in your personal life. Many people use Claude, but the way they use it is like using this iPhone. It has thousand features, but most of the time people are using it just to make a call or send

[00:46] a message on WhatsApp. So it's like using five features out of thousands. Claude is not just a simple question and answer tool. You can do much more. We are going to cover use cases such as building rent versus buy simulator. We

[01:02] will analyze documents like my car a repair quote. We will also connect it with Google Drive and we will analyze the expenses. We will create categories.

[01:14] We will create the expense summary sheet. Then we will use a feature called co-work where we will connect Claude desktop application with a local folder and we will automate the presentation. We'll automate basically creating

[01:30] PowerPoint presentations using co-work. Then we will look at the agentic use case where we are going to audit a website. At the end, we will build a website. At the end, we will build a very useful Chrome extension using

[01:44] Claude code. Don't worry, you don't have to know coding. We will still build this useful Chrome extension. Now this is going to be a hands-on tutorial, so it is going to take time. It is not like those videos that you find in YouTube

[01:58] where they say learn Claude in 15 minutes. Folks, if you want to learn it properly, 15 minutes is not an enough time. So, if you're looking for a shortcut way of learning Claude, this is not the right video for you. All right,

[02:11] let's get started. As a first step, we'll create an account in Claude. Just Google Claude and go to claude.ai website. Here you can log in using Google or you can create your individual credential using your personal email ID.

[02:25] Once you enter, you will see chat GPT-like interface. Now, you are on a free plan. Free plan includes a lot of good things, but once you use it for some time and if you feel that you need more quota or more usage, then you can

[02:42] upgrade to pro plan, which is only $17 a month. This diagram shows you the leverage that you can get by getting on a paid version. So, here each dot

[02:54] represents 3.2 million people. Gray dots are the people who do not use AI, around are the people who do not use AI, around 6.8 billion people on this earth. Then the green dots are the people who use free version of chat GPT, Claude, etc.,

[03:09] 1.3 billion people. Yellow dots are the people who have paid version. So, when you buy a paid version, you get into this 0.03% this 0.03% top 0.3% of the world population who is

[03:23] leveraging AI for productivity and for building amazing things. building amazing things. And if you use AI on a regular basis, if you're using it to code and build things, you will come in this red dot,

[03:37] which is 0.04% of population, around 2 to 5 million people. The general guideline is start with three and the moment you find it valuable, upgrade it to pro version. It is $17 a month. You can spend it for 1 month. Next month, if

[03:51] you don't like it, just cancel it. Let's try our first use case of trip planning. The co-founder of Code Basics, Sam Anand, just sent me this message on WhatsApp about my upcoming trip to India. Now, I have lot of like digestive

[04:06] problems and I can't tolerate spicy food. So, I'm always on a lookout for non-spicy, mild food options. So, in this Claude interface, which looks like ChatGPT, I'm going to give you this prompt that I'm

[04:21] traveling to Bengaluru. This is my date. Suggest me some best vegetarian restaurants which offers less spicy and tasty food. AI will generate better answers if you can give better context. It's like

[04:37] garbage in, garbage out, okay? So, you have to give your requirement. I just said this is my travel date. I need a less spicy and tasty food. And I also gave some of the restaurants which I like. For example, I was in Bengaluru

[04:50] last time, I like Satvam. And I'm supplying all that information to Claude here. And look at this, it generated the answer. You know, it is suggesting this place. Uh I have also given my location. So,

[05:03] I'm planning to stay in HSR Layout. And it will do uh Google search. It will call bunch of tools. And it will find locations which are near to the place where I'm planning to stay. Here, it is using this Opus 4.8 model, which is a

[05:20] very powerful, capable model. You can go to more model and select one of these. You can also look at the afford where if you are giving it a highly complex task, then maybe you can pick uh the, you

[05:34] know, this option, high option, etc. You can also use this option to give voice Suggest some options which are open after 10:00 p.m. See? This will save you a lot of time. I use this audio option all the time and I

[05:49] use this audio option all the time and I just talk to it as if some expert PhD guy is sitting next to me and it is obeying my commands. See, so what it is doing is it is going to internet finding information on the store hours etc. And

[06:02] it will find me the places which match my criteria. Let's do a quick overview of the options which are available. If you open this left panel, you will see a history of all your past chats. So you can go to any past chat and you can

[06:17] continue your conversation. You also have these options of code. So this is something you use to write code. Then you have something called co-work. We'll look into it more and you can also create these projects. So here, let's

[06:34] say I I'm buying a solar panel and I have some details. So I can create project. I can upload my electricity bills, all the details. This is similar to GPT projects. And then when I ask any question, it will have that context.

[06:49] See, this is very important. AI needs to have enough context of your situation. Then only it can produce high-quality answers. Using this plus option, you can add files. So if you want to do any file analysis, just say plus. You can also

[07:04] add connectors and we are going to look at all these options in a bit. While working with AI tools, you need to protect your privacy and there are couple of options that you need to look at. So here, go to settings, click on

[07:18] privacy. There, I have disabled both of these options. Look at this. Help improve our AI model. Allow the use of chats and coding session to train and improve Anthropic models, which means whatever

[07:32] data you are giving it, it will use it to improve their models. I don't want them to use my data for their model improvement. So I have disabled it, okay? And I suggest you do that as well.

[07:46] okay, you are not supplying any personal information and you are okay with this Frontier Labs training their models on your data. You can explore memory preferences here by clicking on manage. Also, remember you should not give your

[08:01] bank password, all your personal details to this AI because your data is not staying local. It is going to cloud server and Tropic folks will have access to your data and you don't know what's going to happen to it.

[08:15] So, never share your confidential information like banking passwords, usernames, or any other thing where you feel that you shouldn't be sharing that information with others. The next use case is car repair evaluation. This is a

[08:30] real use case. My Toyota Camry had issues and I went to a garage and this person gave me this price quote. Now, I'm not sure if this guy is overcharging and is this a fair pricing or not, okay? So, I took help of Claude. What I did is

[08:47] See, I had my personal information, so I blacken that out using this tool called blacken that out using this tool called PDF 24. So, if you go to this tool, it will look something like this and it's a free tool, by the way. And I just masked

[09:02] my personal information and this report I'm now going to give to Claude. Now, whenever I'm working on a new problem, I always create new chat. That way there is a clean history of it and it doesn't assume things based on your past

[09:16] conversation. So, here you can just say plus add files or you can just drag and drop this file here, so it has access to this file. Now, I will give it a prompt.

[09:28] Just give this kind of prompt, okay? I live in NJ and this is the estimate I got. Tell me if he's overcharging. For doing analysis, it is going to multiple websites. So, LLM, large language model, is a brain and it is using all these

[09:41] website, taking information on latest parts, you know, how much they cost in the place where I live, and then based on that, it will do its analysis. My analysis is now ready and it is telling me that the

[09:56] ready and it is telling me that the price for TPMS sensor is high, right? Price for CV axle is very high, high, high. This way now I know that I need to go to two different other garages and get the price estimate for this work.

[10:10] You can attach more than one files as well. The other day I wanted to do my electricity bill analysis and I had these two bills. This one was from June, this was from May. I attached them and I just said that what is the percentage

[10:24] difference in my bill in the subsequent month and it told me exactly that it's 10.6% up. It will tell you the dollar difference as well. Not only it can do descriptive analysis, it can do

[10:37] descriptive analysis, it can do prescriptive analysis as well, which is look at the bills, give me three practical tips to reduce my electricity bill. And see, it is giving me some pretty effective

[10:50] tips that I can use to cut down on my electricity cost. I feel that if you use AI to optimize your life, that $17 a month you're spending is going to be nothing. You will get much more return on that investment. By the way, this is

[11:05] not a paid promotion from Anthropic. I personally use this and find it very valuable, hence I'm suggesting this. The next use case is pretty interesting. It is applicable to all of us. Whenever we are buying a home, we have this

[11:18] confusion, should we buy it or should we rent it? So, why don't we create an interactive dashboard in Claude which helps us with that decision making? The prompt that I have is build me HTML simulation for this scenario where the

[11:34] flat price is three crore. Okay, prices are crazy. And the interest rate is 8% so you'll be paying a lot of interest on your loan. your loan. versus if you rent it, what's the

[11:48] trade-off? Okay, so I'm going to give you all these prompts and all the assets will highly recommend when you're going through this tutorial, please try these things out. You will realize the power of this AI tools first hand. If I'm not

[12:02] buying a home, I will have some capital which I can invest in land, mutual funds, etc. And you will be growing your money. Whereas if you're buying the home, you are paying hefty down payment, you're paying huge mortgage interest

[12:18] rate every single month. So is that worth it? Will I get enough appreciation on my property? So this HTML tool will help answer all these questions. It created this HTML file which I double clicked, opened in

[12:33] file which I double clicked, opened in my browser, and look at this. Here I can increase or decrease my flat price. So let's see if I'm buying a two crore flat with a loan amount of two crore. Let's say loan amount will be

[12:45] less because you'll be making some down payment. Okay, so let's say this is your loan amount. My interest rate is Let's say you get a good deal, right? So here

[12:58] for these parameters uh rent and invest wins. Okay, they're suggesting don't buy this flat. Instead, rent a home and whatever money you have, You can play with the property appreciation as well. See here, let's

[13:14] goes up. Let's say you're buying a property which is uh getting appreciated by a huge amount. In that case, maybe buying wins. See, buying wins by this much, okay? Because

[13:28] your property is getting crazy appreciation. Uh you can look at all these parameters. Look at this tool, folks. Like before AI, this would have taken, if you're a programmer, it would have taken you few months. And here,

[13:42] Claude built it in few minutes. How good is that? You can also print and save PDF for a given scenario, which contains all these specific parameters. Here it is showing you the network. So, let's say if you rent versus buying, then your net

[13:58] worth will be this much at the end of, let's say, 20-year time period. See, here it is showing you the growth of your net worth if you buy versus rent. So, if you rent, your net worth, look at the green line, will grow faster. And

[14:12] portfolio, okay? Here you can look at all these parameters. Please take a look at it, folks. This is so useful. People should use Claude to build this kind of interactive tools whenever they're buying something, and that will help

[14:28] them make a better financial decision. If you have been using other AI tools such as ChatGPT, etc., you might have important context of your past conversation in that tool. You can easily import that in Claude by

[14:41] following these steps. All you have to do is click on settings, then go to privacy, click on manage memory, and import memory from other AI providers. And you will copy this particular prompt that

[14:57] Claude is giving you to, let's say, ChatGPT. ChatGPT will produce some question to ChatGPT, it will produce some answer. You take that answer, and you copy paste here. See, paste results here. And that's it. Whatever context

[15:12] here. And that's it. Whatever context you had from ChatGPT, now you have it here in Claude as well. And this will be helpful because now, through the past context, you can know your preferences, your past background, and it will have a

[15:25] general understanding of working with you in a better way. The next use case is connecting Claude with Google Drive. This is going to give Claude superpowers because now you're giving it access to your private organizational data, and it

[15:40] will be able to save you a lot of time on tedious work. For example, I have all my team expenses for my team in this particular drive. Now, what they do is like they go for coffee, they take a cab, they will take a picture, they will

[15:54] upload it to this drive. I need to organize this folder, okay? So, let me show you the same folder here locally, and you can see all these images. This one is a bill from Amazon. This one is some food bill. This one right here is a

[16:09] bill on some office supplies. My end goal is to convert this one directory, which has all the dump, all the garbage, into this kind of organization, where you have folders for specific category. This contains the bills only for food

[16:26] And it has dates uh as a prefix. So, if you are doing any analysis, it is useful, right? So, it has dates. It has all other categories as well, groceries,

[16:40] office supplies. Not only that, it will have an expense summary file. It would be beautiful if Claude can generate it, right? Where I have date by date right? Where I have date by date expenses by vendor, category, amount,

[16:53] file name, and so on. Usually, if you want to build something like this, it's going to take a lot of time, but I will show you how you can do this with Claude very quickly. I have started a new chat in Claude, and I will click on plus icon

[17:06] here, and here you can find all the connectors. So, you can just say manage connectors. I already have some connectors. You will probably see this blank. So let's add connector for Google, okay? So plus browse connectors.

[17:23] And here you see the connectors from all major software providers. You can connect with Slack, Notion, Google Drive, you name it. Here we are going to choose Google Drive. Click on connect. I will use my Google account here. I had

[17:37] issues with my previous account, so I connected with a different Google account and it got connected successfully. And here you are seeing all the permissions. There is read tools and write tools. Meaning you are

[17:50] allowing Claude to read content from your Google Drive. Right tool means your Google Drive. Right tool means you're allowing Claude to copy, create files on your Google Drive and so on. So here this says needs approval, this says

[18:05] always allow. So I'm going to say always allow for all the read-only tools, which is Claude can go to your Google Drive, it can read all the content. Now my Google Drive for the different account looks something like this, okay? This is

[18:21] my company's folders. So your courses, media kit, your team expenses, and team expenses is this flat list of screenshots. Not just screenshots, it's PDF as well, see? There are PDF files. Now in PDF file you can copy-paste the

[18:37] data, but if you have image you can't copy-paste, right? Image is image. And you will see that Claude will be able to grab or extract information from images as well. All right, so here is how my connection looks. Connection is okay. I

[18:52] connection looks. Connection is okay. I can go back and here, see? Just verify, see this you can say it. And temporarily, if you want to disable it, you can do this as well.

[19:05] Now you can ask question, "Give me the list of folders in my Google Drive." And it should give me this list, that that I have these five folders. I will say always allow. I gave that setting already when I was configuring, but it

[19:20] asked me again. I don't know why. You see, it is giving me the expected information. This this will verify that your connection is okay. The next step will be this. Look at all my files in this folder, organize them into sub

[19:33] this folder, organize them into sub folders by expense type, rename each file with clear format date, vendor, amount. Let's go. While this is working on it, one thing I realized is, you know, I need to give it a right

[19:48] permission as well. So, I will go to cloud.ai, open a new instance here in a new browser tab, then go to connectors, manage connectors, go to Google Drive, and here I will just say always allow, okay? Even

[20:02] if you didn't have that, it will ask you for the permission, but I want to make you're doing complex analysis, you might need a pro plan because with a free plan, there will be some daily usage limit. Here it is saying that the driver

[20:15] connector can copy and create because see, it has copy and create access. It cannot move or delete the files. So, how should we proceed? See, it's giving you the option. I will just say renamed copies plus summary sheet. So, let's

[20:29] click on that. It has done some work and some of the files are organized, so I would like to go and take a look at it. And look at this. You see now these beautiful folders. You still see these old files because we don't have the

[20:42] old files because we don't have the access to delete files on Google Drive, but it just made copies. So, if you look at food, you see all these food receipts. See, Subway, Biryani, Domino's. Look at this. If you had a

[20:56] person sitting and organizing these expenses, it would have taken them hours, but Claude can do this accurately in minutes. It also put the date as a

[21:08] prefix so that when you are processing these files later on, you know for which date the expense is for. At some point it stopped because when it was doing thinking, it went above its limit for the tokens

[21:21] and then it said, "Okay, should I continue?" I said continue and it finished the remaining work. It is in fact doing the work right now, but at the end you're going to get all these category folders with the appropriate

[21:33] files in it. It took very long and finally it is done. The task was finally it is done. The task was complex. It had to read images, PDFs, and so on and I believe that is the reason it took long time. In fact, I ran

[21:46] out of my limits and I had to add some usage credits. When you are trying this thing out, what you can do is you can reduce number of files. Okay, so instead of so many files, just have three, four files. See the client meeting folder

[22:00] contains all the client meeting related files, then food, then you have office supplies and so on. Not only that, it also created the expense summary. It

[22:12] created on its own. If it doesn't, you can give another prompt saying that, "Can you create expense summary Excel file?" Here it created it and the file looks something like this. Look at this. You would have an analyst, a person

[22:27] sitting in your office doing this work manually, but Claude did this on its own. So there are categories: food, travel, office supplies, and so on. And there are clear columns: the vendor, date, amount, original file name, and so

[22:43] on. The field of AI is changing very fast. If you want to stay up-to-date with the latest tools and approaches and if you want to improve productivity in your work as well as benefit in your personal life, then we have this

[22:58] CodeBasics AI Pro subscription. It's a monthly subscription where every two week we have live build workshops where we will take a real problem and solve it with you on a live Zoom session. You will also get weekly AI hype versus

[23:14] will also get weekly AI hype versus reality digest. Then there are monthly hands-on challenges and our students are building amazing things. For example, she's a homemaker. She's now building a real tool personal finance tracker. Uh

[23:28] game for kids and so on. People are building amazing things with these powerful tools and gaining benefit in their personal as well as work life. If

[23:40] you want to know more, we are going to put a link in a video description below. Next important thing I want to talk about is skill. Just like how we as humans have skills, right? Somebody will have a skill of swimming. I have a skill

[23:53] have a skill of swimming. I have a skill of writing Python code, etc. Similarly, you can have skills in AI as well. Let me show you by creating a skill in Claude. So see here, whatever we did to organize our folders and create a

[24:08] spreadsheet for expense summary, we can save that thing as a skill. You can just ask Claude, you know, in a natural language, just ask that whatever you did so far, save that as a skill. It took few minutes and finally it has built the

[24:22] skill. Skill is nothing but an MD file. See, skill.md. And it's just a markdown file which has a description that can help LLM perform this task next time. So next time I have similar task, I can give this skill and

[24:40] it will perform the task for me. If you read through it, it is just a detailed information of the task that was performed and this will be useful in terms of clarifying our expectations to an LLM. It also uses some Python file

[24:58] for doing that categorization, so it created that as well. Now, I can say save skill and see it has saved the skill. When I clicked on manage, I will see all my skills here. So, right now I have expense receipt organizer as one of

[25:16] the skill here. See, you can uh read through the MD file and look at all other details here. Next time you want to use a skill, just say {slash} name of the skill. So, I will say expense receipt organizer. And let's say

[25:32] now I have connected my OneDrive, and in my OneDrive, I have a special folder where I want LLM to do this expense categorization. In that case, I can say that in my OneDrive folder. Obviously, you have to

[25:49] go and connect your OneDrive using the connectors the same way that we did with Google Drive, and you can say that in folder folder you know, you just specify the path. Uh

[26:03] perform uh categorization of receipts, something like that. And it will use the saved information from this skill to perform that task. Next thing I would like to discuss is plugin. Let's

[26:18] say you are an AI-enabled HR person, and you have created this HR person, and you have created this skill called offer letter, okay? So, that's a skill.md file, okay? This is the first skill

[26:32] that you have created. And you have created another skill called onboarding plan. So, let's say onboarding plan is the second skill that you have created

[26:44] as an HR manager who is very good with AI. Usually, you use these skills along AI. Usually, you use these skills along with one HR database that you have in your company. Okay, so this is nothing but an HR database.

[27:00] Plugin is [clears throat] nothing but a bundle of all this. So, let's say you want to bundle these three things together because you you work with these two skills along with this database all the time. So, now you can create

[27:13] something called plugin. So, plugin is nothing but skills plus integration. Here is a HR database integration. This is more like the extensions that you get in VS Code. If you have used Visual Studio Code etc. You know, you

[27:28] plugins or let's say extensions. For example, Chrome extensions, right? You have used that. So, plugin is short of like that. It is skill plus some integrations. To create a plugin, you can just say bundle

[27:43] the skill I just created with Google Drive connector into a plugin. Save and install this plugin. See, now it is writing this into a plugin. You see? So, I can wire it into a plugin properly. It has created the plugin and

[27:59] now you can install it. See, that's the install step. It is like installing Chrome extension. So, you are saying save plugin and it has saved it to this particular folder that you see here. And if you want to see all the plugins, see

[28:13] here in the connectors. You see personal plugins. Now, you are seeing this expense receipts plugin. Now, in order to use this plugin, all you would do is say, "Can you organize the expense receipts into my Google folder into

[28:27] spreadsheet?" You are not giving detailed instructions, but what Claude will do, it will look at existing plugins and skills and it will use the

[28:39] one which is appropriate for this task. So, see I have this team expenses with a So, see I have this team expenses with a flat list of files. Now, when I give this particular command, that plugin will kick off

[28:51] and it will do the organization. So, I'm just going to fast forward this. See, it created that expense summary file and everything. You can go to claud.com plugins and find all the plugins in the marketplace. See, if you want to do

[29:06] front-end design, there is this plugin. Contact 7, there is this plugin. So, these are the things that you can install in your Claude code and even get benefit of the work that other people have done to build these plugins. The

[29:21] cleanest way to remember this is a connector is a door to an app. A skill is knowing how to do a task, whereas plugging is a kit of doors and skills.

[29:35] Now, folks, there are three ways we can run Claude. Until now, we have run Claude in a web browser. Okay, so that is the first way. It is very quick and you can run it anywhere. The second one is to download Claude desktop

[29:49] application on your computer and have it work with the files on your computer. The third one is a mobile app. So, if you go to your phone, you can download Claude app and you can do the work while you are on the go. To install the

[30:04] desktop version, you can go to claud.ai, click on download desktop app and just save this file and just run the installation, your usual common sense steps. So, folks, here, this is how the desktop version looks. If you look at

[30:21] the left panel, it has chat, it has co-work, and it has code. Code is something you will use to write your code. Here, we are going to use co-work, code. Here, we are going to use co-work, okay? So, the way co-work works is it

[30:35] will help you connect Claude with the local folders on your computer. For example, the expense use case that we discussed before, let's say the team expenses are not on Google Drive, but

[30:50] they are on local drive. So, in my C data folder, I have team expenses. And just to save on the token cost, okay? So, let's connect this folder here. So, you will go here and say choose a folder.

[31:11] And then select your folder where you have your expenses. Click select. Allow. And here you can select the mode. Ask before acting, or you can say act without asking. Okay, so I'll just select the act without asking.

[31:28] Now, see the expense skill that we created earlier, you'll not find here because that is saved on your online account, right? This is the local desktop version. So, here I'm going to give the prompt. What is my prompt? Look

[31:44] at all the files, whatever that prompt was. When I do that, it is going to work on that particular folder and organize them. See, it found nine files. On the progress. See, these are the list of

[31:59] tasks, and it will check those tasks as they are completed. Look at this. Ta-da! It has created the expense summary sheet, and also organized things into sheet, and also organized things into multiple folders. See, client meeting,

[32:13] food, office supplies, and so on. It did not delete the original files, but you can give it the instruction. You can say that move the files, you know, do not keep the duplicates. And the expense summary file, this one, is present here.

[32:29] You have all the expenses. Now, you can do some Excel formulas, group by, and you you do aggregations, and so on. The next task that we are going to do using Co-work is automating PowerPoint presentations. When I'm teaching on this

[32:44] channel Code Basics, I'm teaching through PowerPoint presentations. And I have a specific way of teaching, specific style. So, I would like Co-work to automate creating this PowerPoint presentations for me, all right? And

[32:59] this is a very valuable million dollar skills that you can develop. Now, I have certain teachings principles. For example, uh I treat my learners as if they're wounded with confusion and distractions.

[33:13] wounded with confusion and distractions. So, I take care of them really well while I'm teaching. And you might have experienced that through Code Basics videos. And when you see any presentation, my uh screen is minimal.

[33:27] It doesn't have too many elements. And even if it has too many elements, they will appear one by one. I never throw a lot of text, lot of visual elements at once at students, otherwise they get confused. So, that is one principle that

[33:42] I follow. So, I have written that principle here. Okay, see? Don't show too many things on the screen, use multiple slides. If they all need to be on one slide, then animate. Now, these are the kind of instructions that I give

[33:54] to my other co-instructors. Right now, we have AI engineering and AI PM cohorts running on where there are other instructors who are teaching. And I will give the same instructions to them so that they can follow my teaching slides.

[34:09] Okay, do not cut or swallow words towards the end, engagement, whatever. See, I have laid down my teaching principles into one text file. These are the teaching principles. But when it comes to PPTs specifically, how

[34:25] to animate, I have different set of principles which I have mentioned in this file, how to animate.txt. Treat each box and its text as a single unit. Okay? See, keep the slide title heading visible, use bigger fonts for

[34:41] diagram step-by-step flows, let me decide the order things show up so I can reveal one piece at a time. You can read through this. These are very good principles that I have learned over past 10 years doing YouTube. And I have

[34:54] created these two text files. So, now I have this folder, you know, in data I have this folder called PPT creation where I have this presentation skills which has these two files, Art of Teaching and How to Animate. It also has

[35:09] one of the slides that we have created previously for one of the courses, okay? So, let me show you these slides. So, these slides are available in our time management course, okay? The same exist slides. And these are created using the

[35:23] same principles which are outlined in this. So, these are the principles and this is one example, you know, my example PPT which I have created in past. Now, I'm going to give this to Claude and ask it to create presentation

[35:38] for a new topic. So, let's do that, okay? So, we are going to go to Claude okay? So, we are going to go to Claude and click work and just say new task, and click work and just say new task, okay? And here connect it to that PPT

[35:51] folder. So, I will say C data PPT creation presentation skill, okay? So, let's connect and just say always allow. I'm saying using the guidelines given in these two files, right? Art of Teaching

[36:06] these two files, right? Art of Teaching CB principles .txt, this is one file. And then how to animate, this is another file. I want you to create three slides presentation for topic Agent TK, use

[36:19] time management and deep focus. This particular PPT, you know, this PPT use it as a sample reference. After some time, I see my slides and looks pretty promising because if you look at my sample that I gave, the background, the

[36:35] sample that I gave, the background, the colors, the style is adhering to our brand guidelines, okay? So, creating PowerPoint presentation becomes so easy

[36:47] if you're using Claude Cowork. In your work, if you're creating any PPTs, you will be using some template from your company, okay? And you'll be using some kind of style, some kind of approach for creating presentations. You can put all

[37:01] of that into text file, and you can guide Claude to do it, and and in the future, it will do that. Not only this, you can create a skill out of this. So, let me create a skill. I can say that use these two files, right? this.txt

[37:18] and this.txt to understand how I want deck to be created, and create the skill out of it. So, let's go and create a skill. So that

[37:30] in the future, I can just say {slash} my skill and give the topic name, and that's it. It took some time, but finally, the skill is created, okay? Now, I will say save skill, and see the skill is saved. So, when I go to manage,

[37:46] I will see those skills. So, here, CB presentation is the skill. Okay, that's the name of the skill. I think you can rename it as well. And there is this skill.md file, which contains the description of, you know,

[38:00] what kind of background you need to create, tile, foreground, all that information. Now, I can go to regular chat. So, see, I'm in Cowork right now, okay? Where I have connected

[38:14] it with the folder, but I can go in a regular chat. This is my regular chat. Here, I'm going to click on a new chat. Here I can say CV presentation and just say

[38:27] and just say Okay, create presentation for context engineering. And then it will use that particular skill because I'm referring to that skill, okay? So, it's going to

[38:42] create see, reading the CV presentation skill, it will read all the details on fonts, background color, my preference on animation and so on and create a presentation on a new topic context

[38:56] engineering. If you want to brainstorm with Claude on the topic first and build a content outline, that is always a good idea. But, I want to keep things quick, so I'm saying, "Okay, you just build a presentation on this." By the way, I

[39:09] changed my topic. I I'm now saying create three slide presentation on what is context engineering. I want to limit number of slides so that it can be done faster. Finally, this is done. See, it created this three slide presentation

[39:23] using the theme, background, style, approach that I have given while approach that I have given while creating the skill. So, for me as an educator now, it is so easy, right? I can feed my content, I can brainstorm

[39:38] ideas, and I can just say CV presentation topic. Let's say I want to presentation topic. Let's say I want to explain convolutional neural network, you know, neural network. And so on, create like 10 slides

[39:52] presentation, something like that, and it will do it for me using my style. Next up, let's do the audit of Code Basics Foundation website. So, this is organization

[40:07] where we are involved in various initiatives, mainly related to kids education. So, we have our team members Manjunath, Hemanand, Naveen, they are participating in different activities. This website was created by Manisha, who

[40:24] is looking at this foundation activities in our team. She is not a technical person. She used live coding to build the website, and the website looks good, but I want to do an audit of this from the UX standpoint. I can use Claude

[40:40] the UX standpoint. I can use Claude here. I'm going to create a new folder here. I'm going to create a new folder called CB Foundation website audit. And I will connect that with my Co-work.

[40:53] So, let's click on this and click on Co-work. And in the Co-work, you will say create a new task. Okay, so the new task is created. And in the new task, you will choose the

[41:08] different folder. So, see data and in that CB website audit. Okay, always allow. And make sure you are unchecking these two folders,

[41:23] okay? Looks like you can attach multiple folders. I want only one folder to be attached to this workspace, and then I will give this particular prompt. So, the prompt is look at this website and do a UX audit. Scroll through the full

[41:40] page and take screenshots. This is very important. Usually, it can read the content of the website, the text content, and do analysis, but I don't want only text content. I want it to take the screenshots of the website,

[41:56] then resize it to mobile width, and screenshot again to check a responsiveness. If you have done website development, you will know about this responsive design, okay? I want to make sure that it has responsive design. I

[42:11] want it to test the key interactions, hover status on mouse button donations as if you are having a QA person in your team and they're doing all this audit. Note anything visually off, spacing, etc. Okay? So, let's go. And the output

[42:28] format here is give me findings with the screenshots as evidence in a Word doc. So, what I'm expecting it to do is now it will follow all these instructions, go to the website, take the screenshot, do the analysis, and it will create a

[42:43] nice summary report in a Word file. Here it is asking me to install Chrome it is asking me to install Chrome extension for Cloud. Let's do that. In Google's, type Chrome extensions. So, click on this particular link. Search

[42:59] Enter. Click on this. Add to Chrome. So, this Click on this. Add to Chrome. So, this is going to add Cloud's extension in my Chrome uh browser. See? You can see it's now

[43:13] added. I can say authorize, so I'm authorizing it just in case if I need to use it. Now, let's say it is now installed. Because I think it probably need that to take screenshots and to

[43:29] grab the information. Look at this. It started taking pictures, looks like. I think not pictures, but it is opening Cloud in Chrome. And here you are seeing the progress of all the steps. If you notice its activity, you'll see it is

[43:44] notice its activity, you'll see it is doing a detailed work. For example, the get involved button appears to not change visually when hover. See? It took a screenshot of that. And it did that analysis. So, it is

[43:57] going step-by-step and doing this analysis. Now, of course, it is going to use a lot of tokens. If you want to limit the scope, maybe you can say that limit the scope, maybe you can say that do the audit and find only uh three

[44:12] items that I need to change and then stop. So, that it is not going through the whole website and doing the holistic review, okay? This is just to save tokens, but otherwise, if you look at real life, for example, this is a very

[44:26] real life, for example, this is a very uh real example. It is our website and as a founder of Codecademy, I and my team want to do the audit to make sure it is adhering to all the UX principles. So, this is something we will do for

[44:40] real, okay? We'll do the holistic review. But, for you, while you're practicing, you can just say, "Find first three issues and then stop." After first three issues and then stop." After 20 minutes or so, finally, my website

[44:54] report is ready. If I check my directory, I will see this nice audit file. Look at this. This is as if some professional has built it. It has two critical issues, three high issues, four medium issues, and so on.

[45:08] I'm most interested in critical issues. So, let me check maybe this one. So, let me check maybe this one. Rightmost volunteer card clipped. Okay, that sounds like a problem. So, when I go to the website, what I notice is in

[45:23] the volunteer section, see, when I have volunteer scrolling, the three volunteers are good, but the fourth one is getting clipped, see? I have three, but you see this little uh card is getting clipped. This is a

[45:38] problem. Actually, you should have a view where you can see all the four volunteers clearly, okay? And then you have other issues, for example, here. have other issues, for example, here. Let me check the very first one, which

[45:53] is get involved CTA. So, get involved button doesn't have any CTA. It doesn't change visually. So, if I go to website, here, see it changes the color, but it is not like this. See here, things kind of move and there is a visible CTA. It's

[46:10] not that apparent and get involved is probably the most important button on this website. Because it's a foundation website, we are looking for volunteer, website, we are looking for volunteer, donations and so on. And nothing much is

[46:22] happening. So, from UX standpoint, it's uh major issue. All right? So, that's all that we had for the website audit part. Let's review whatever we have done part. Let's review whatever we have done so far. We started with normal Claude

[46:36] planning. [music] We gave it files for our analysis of bills. We built interactive tool for our buy versus rent decision making. Then we

[46:48] went to step number two, which is using Cowork, where we connected Claude with Google Drive, with local folders. We did some file organization. We did analytics

[47:00] for our expenses categories. [music] Now, we are going to enter the Ghost Rider Builder zone. Here, we will be building a real software tool. Let me

[47:12] show you how it looks. Many times when I'm going to different websites, I want to take notes. For example, here is a nice quote from Ramakrishna Paramahamsa. nice quote from Ramakrishna Paramahamsa. What if I can just highlight this quote

[47:26] and save it? So, see, what if I get this kind of save highlight button, where when I save it, you can say you can save it as a note. Let's say Ramakrishna's quote, okay? I can hit save.

[47:40] And then I can make another note for this news that you know, this noble uh prize winner person is now joining Anthropic. So, I can save it as AI news,

[47:54] Anthropic. So, I can save it as AI news, okay? And the third one is dear Grogu. I can say baby Yoda's intro. And then all these see? So, this is the extension, folks, that

[48:09] we are going to build, okay? So, we'll write code using Cloud Code. It will build this extension. We will import it in Chrome. And eventually,

[48:21] you'll be able to add all these notes, and you will have this kind of dashboard where you can see the notes. Not only that, if you want to visit that page, you click here, you'll be able to visit it. This sounds amazing, right? It

[48:34] requires real software skills, so many years of experience. And now, with Cloud Code, you all will build this. I will create a new folder to save the code for

[48:46] this note taker, okay? So, I will say it note taker extension. And then, I will go to my Cloud Code. So, see here, you have to select this third part. Just click on code,

[49:00] okay? And then, connect your folder, which is C data note taker extension. Okay, so

[49:12] trust the workspace. Now, you can uh create something called a PRD. So, PRD is nothing but product requirement

[49:25] document. Whenever software teams build software, they put the requirements in a software, they put the requirements in a document called PRD, where they have written the purpose of the software tool, who will be the users, technology

[49:40] tool, who will be the users, technology stack, any limitations, considerations, and so on. So, we will use Cloud Code to generate PRD first, and then using that PRD, we will generate the software. So, the PRD generation prompt may look

[49:55] something like this. I want to build a Chrome extension that helps in capturing the highlights in any of the websites, store it, create a PRD in markdown format. Markdown is basically

[50:07] it's a very lightweight format with some special tags for headings, uh normal text, and so on. Okay? See, it created the PRD. Let's look at this PRD file.

[50:20] So, if you click on it, it looks something like this. So, it has a summary, the tool's purpose, you know, problem statement, goal, non-goals, which means you don't want to cover these things which are advanced,

[50:35] and target users, key use cases. If you want to edit it, you can edit it. You can open it in Notepad or things like Visual Studio Code. See, Visual Studio Visual Studio Code. See, Visual Studio Code is a free code editor. And here,

[50:49] I can open that particular file. Note Taker extension was our folder, right? So, it created PRD.md file. MD is extension. And when you look at the file, it looks

[51:04] like this. It has all these headings. See, headings is like two hash, three hash, etc. And when you see it in a preview mode, it looks like this. You get this nice formatting done for you. Now, if you want to edit it

[51:20] just like simple text file, you can add it. You can add goal, remove goal. You can update it, okay? We're going to keep this as is. And here in the goals, I would like to add one point uh where I will say, "Build

[51:36] a dashboard where all the highlights can be seen." Okay? So, add and update the Okay? So, add and update the requirements. Now, you can give this PRD

[51:51] to Claude code and ask it to generate the code. See, it is already asking, "Should I scaffold the actual extension?" So, we can say go ahead and build

[52:03] the actual extension now. And it's going to take some time to write the code. You can take a coffee break meanwhile. It is done writing the code. It not only created code, it ran

[52:16] it internally and made sure things are okay. It also gave instructions on how to import this extension in Chrome. Okay, so these are the instructions. We are going to follow

[52:30] these instructions, okay? And before that, let's look at our directory. I I'm curious. See, in the note taker, I think note taker here, you see all this code. Source, we have all this beautiful code that is returned. So, let us follow now

[52:48] that is returned. So, let us follow now the instructions given by Claude here. So, we can go to extensions here. Okay? You can go from here also, see? If you go to extensions and if you say manage

[53:02] extension, you will land here. Click on developer mode because this is in a development stage, we want to enable developer mode and then click on load developer mode and then click on load unpacked and select that particular

[53:16] folder where we have our source code. Okay, so note taker extension is the folder where we have our source code. Select the folder. When you do that, it says extension loaded, okay? So, now when you See, this

[53:32] is the extension basically that it created. Now, I'm going to open my Baby Yoda's website. I will highlight, see? In earlier, when I showed the extension,

[53:46] uh it was created by one of my team members, so look and feel was different. So, this is the thing with generative AI. It look and feel might differ, and if you want consistent look and feel, then you can specify that in your PRD,

[54:01] okay? So, I will just say save highlight, okay? And note is Baby Yoda's intro. Okay, highlight is saved. I go here,

[54:16] look at this, and I can say open dashboard, and see in the dashboard, I am seeing that particular thing. Go to another website, save the quote, okay? So, you can just say save the quote.

[54:33] I have saved it. Highlight saved. When I refresh the I have both the quotes. All right, you can export all of them as a CSV file. You can also import. You can delete all. You can add, edit, see? It implemented

[54:49] so many other features which I did not even intend to implement. So, it is very smart in that regard. If you want to pin it, you can just say pin, and now see it appears here. So, you are on any website now, and you want to quickly

[55:05] any website now, and you want to quickly grab the news, okay? So, let's say uh news on volatility, okay? So, you can just say select, save highlight,

[55:18] save, done, and here you can see the quick preview, or you open the dashboard, and you will see it. Isn't this powerful, folks? Now, you all can build software. You don't need

[55:32] to go to uh the detailed length of training. Of course, if you have engineering skills, you will be able to build large-scale production systems, but to build this kind of small tools, now you can do on

[55:45] your own using Cloud Code. Now comes the most exciting and interesting part of this entire tutorial, which is the take-home exercise that you will work on to build this five real-time use cases. Learning technology is like learning

[56:01] volleyball. If you just watch it on TV, if you watch all the videos, you're not going to learn it. You have to go to the ground, you know, take the ball, and ground, you know, take the ball, and start smashing it. These use cases are

[56:14] available in the video description below, where you have things like, you know, you are organizing files and doing analysis into one clean Excel. Uh here are the screenshots. We have given all the hints. Just read through it, folks.

[56:29] You will be amazed at the things you can build which are so useful in your personal life as well as in your work life with Cloud Code. All right, that's all we had for this tutorial, folks. If you liked it, give it a thumbs up. If

[56:44] you have any questions, post in the comment box below. Thank you very much comment box below. Thank you very much for watching.