What Claude Code is

0:05 I have done all these certifications as well. And this video, I have kept it beginner friendly so that even if you do not know C of Cloud or Code, you will still be able to understand everything, do all the hands-on. And by the end of this video, you will be you will become an expert in Cloud Code. And here we will start with some ABCDs of Cloud. Then we will understand what all interfaces we can use for Cloud Code like Cloud Desktop, VS Code extension, Cloud CLI.

0:37 After we understand the core concepts here, then we will understand how we can extend its capabilities using MCP servers, skills, plugins, and hooks. And after we understand this part as well, then towards the end, we will build an end-to-end full stack project where we will see happening that we create some tasks in Jira, Cloud will understand those tasks, complete everything for us, and directly create a pull request in GitHub. All using a single line of prompt that we will give to Cloud.

1:10 So by the end of this video, you will become an expert in Cloud Code and you will be able to work in any project with all the best practices that are used in the industry. I have uploaded all the important files, commands, and other documents in GitHub. I will also be sharing this presentation link as well. And to help you further, I have also prepared a detailed PDF document that we can use as a developer guide whenever we are working on development or even preparing for any interview.

1:46 So the links for all these is available in the description. I hope you are already excited, so without any delay, let's get started. Let's start by understanding what exactly is Cloud Code. So Cloud Code, it is an AI agent that understands your codebase and helps you build, modify, test, and manage entire software using natural language. So putting it in simple terms, it's an AI agent that understands our complete requirement and it can help us from writing new code and new features to making any modification, then testing, and even deployment.

2:32 So end-to-end, whatever we do related to software development, it can do all those tasks for us. And now let's understand some of the core features of Cloud Code. So one is about understanding your project. So it can read our entire codebase. So in in a code we have so many files, sometimes even humans they cannot understand each and everything, but it can understand our entire codebase. And apart from code, it also can understand our complete architecture, the dependency, across entire codebase.

3:12 And then it can also answer project specific questions and make project specific decisions as well. And then, the the core piece or where it started is about writing and modifying code. So it can create new features from scratch. It can refactor existing code, and it can fix any bugs automatically as well. And the next feature is about running development tasks. So apart from just understanding and writing code, it can execute everything as well. It can execute terminal commands, it can run tests, it can even deploy our entire code and even test everything.

4:06 And beyond coding also, it can do so much more, like we can extend its capability and connect to external systems like GitHub, Jira, data bases, all those things we can do. And we can automate complex development workflows as well. So all these things we will understand more when we will go to the demo and do these ourselves. So in a single line if we have to summarize, so think of Cloud Code as an AI software engineer that is working alongside you.

4:44 It is not just an AI code auto-complete tool. It is like a a helper engineer for you, who can do anything for you, you just need to guide it what exactly needs to be done. And now let's understand like how come Cloud Code became so famous and when exactly did it start and at what time like everybody started using it. So that is what I call it as Rise of Cloud Code. So it started in September 2024.

5:18 It was an insider project in Anthropy. So in Anthropic, there was this engineer like Boris Cherny. He started this Cloud Code as an internal project to help with the development related stuff. Then In February 2025, it was like released for research preview mode. So, that researchers, they can like start using it and it was like released as the first CLI-based coding agent. So that people can do research around it and understand what exactly is possible with this.

5:20 And then in May 2025, that is when it became general availability. So what it means is that this time like it was open for entire world to be used. And it was initially released with Claude 4 model families. And then in September 2025, that is what they call it as V2.0. So, there like they extended the ecosystem apart from CLI, they released the VS Code extension, they introduced concepts like checkpointing. And then it became more mature ecosystem with hooks, skills, background tasks, SDKs and all.

6:20 So this was a, I would say major release after the initial general version. And with this major release like it was very easy to use, especially with the VS Code extension and all the other features. So the capabilities, those increased a lot, apart from just using a CLI-based tool. And then came came November 2025. So this is called as the inflection point in the Claude history. So this is when they released this model, Opus 4.5. So the earlier models that were being used with Claude code, those were like efficient.

7:20 But this Opus 4.5, it took that to a very new level. With this model release, we got the real code understanding capabilities and writing the code more efficiently. So that when everybody was using it, they were amazed to see what it can do. So this was the real inflection point where everybody they decided that yes, this is a game changer and it is very helpful. And everybody started using this. And then in January 2026, they released Copilot that shook the entire markets across entire world.

8:00 So with Copilot, we can do so many things automatically and with minimum effort without much background of coding. So it was no more about coding. So with Copilot, we can automate so many things. So that is what was released in January 2026. And by May 2026, so it has been like adopted by so many people across the globe, that it has become the de facto standard of the AI agents. Everybody started following the standards of Claude code.

9:00 Even other companies who were who wanted to build their coding agent, they started to follow Claude code. And everybody like they started using it, they realized like it is the most helpful thing in their development life cycle. And people started seeing its value. And the main point, it is like towards its adoption. So because everybody started using it, so that is why it became more and more famous. And it has now become the core standard, how an AI agent shall work.

10:00 And then other companies also they released their coding agents in the similar formats. So by May of 2026, like the revenue run rate, it rose to 2.5 billion, which was like more than doubling since the even start of 2026. So even in like five months, the revenue rate doubled, which proved like how entire world is using Claude code now. So we can see that in just over a year, Claude code, it evolved from an internal side project into the most influential AI coding platform.

10:20 So that is the rise of Claude code and we can see like how fast it became very famous. And it happened because everybody started using it. And they realized what value is coming out of it. And it was helping every developer in their day-to-day tasks. So now as we know about rise of Claude code, let's understand what others are there in the market and where they stand. So Anthropic, they they are the first one to build this kind of product.

10:20 And then others are also like building similar. It was like I would say natural outcome. It had to become, it had to be there in the market by some company. But Anthropic were the first one to to build it in such a level that it was very useful and helpful for everybody. So they are the market leaders. Then Open AI, who is the natural competitor of Anthropic. So they have also released Codex CLI. So this is also a coding agent from their side.

10:06 Google, they also have released Gemini CLI which is a very similar to Claude code and Codex CLI. And Google, they have also built and released their own IDE like VS Code. They created a fork or kind of a copy of VS Code and they have built their own coding IDE which is called as Anti-gravity. And then there is this Cursor AI also by AnyScale company which is also very famous when it comes to AI based coding agents.

10:38 So, we can see here that the AI coding landscape, it is evolving very rapidly. Multiple companies, they are building different different agentic development tools. So, Anthropic, they started it, but now everybody is trying to catch up. And, and I would say like if we are doing some core basic task, then I have tried also like Codex CLI and Gemini CLI, they can also do a lot many things similar to Claude code. Cursor also can do a lot of things now.

11:16 So, I would say now it is a competitive mode that everybody is in. But Claude, because it is already adapted by everybody, it is still leading. And this is the standard that everybody is following. So, now as we know the core details and hope you are excited that how powerful it is. Then next, let's understand how we can use that power. So, now let's see what are the different ways how we can access and use Claude code.

11:57 So, Claude, it comes up with different options of accessing. So, the easiest one now, this is called as Claude Desktop. This is like best for getting started. This is like a simple like web page kind of application that is that we can install in Windows and Mac and within that application itself, we can write the code and do our core development tasks without opening any IDE or going through CLI. So this is a like a software of Anthropic that they have released, called as Claude Desktop.

12:38 The other option is IDE integration. So, it is like an extension that we can install along with the IDE like integrated development environment that are famous or that are used by developers across the world. So, two of the very famous one are VS Code and IntelliJ. I am familiar with VS Code and this is the one we will also use in this video. And with these, with IDE integration, we can directly work with our favorite IDE.

13:14 There itself, we can edit, review and debug code without leaving the editor. And as people are already familiar with these IDEs, it seems very seamless to use this Claude code there. And the next option, that is CLI. So, I would say this is the core and most powerful way of using Claude code. So, this is where Claude started. They released first as a CLI tool. CLI means like using command lines, we can use this entire Claude code.

13:54 So, this is for power users. So, whatever things generic things we can do using IDE extension. But, if we need to use more power, like using MCP servers or doing some advanced stuff, then we use CLI. And CLI, it provides the complete agentic development experience. This is the I would say the heart of Claude code. So, whatever other things are there, internally they use this CLI tool to do that stuff. They are kind of interfaces. But this is the core interface that is that is used for Claude code and do everything whatever is there.

14:44 And this one, it can work with any repository and any text stack. This is ideal for automation, scripting, and any advanced workflow that we want to build. And then, the fourth way of using Claude code is via SDK. So, SDK, it is like a software development kit that we get. So, this is like something not to let's say help with us, help us like with some coding and that kind of stuff. It is like if we want to build our own application that can like build code and related activities.

15:27 And at the backend, if we want to use Claude code. So, whatever Claude code is actually doing, we want to let's say build a good application and wrapper around it and then let's say sell it or service, serve it or customize it. on our own way then we go for this SDK. So this is powerful if we want to build our own applications and integrate cloud code itself rather than building some code with it then we go for SDK.

15:20 So these are the four options that are there. So we can see it is one coding agent with multiple ways to work with. And just to summarize which one we shall use when, right? So if we, let's say, we are very new, we just want to try out what exactly is cloud code, we can use cloud desktop, it is very easy to use and beginner friendly. But for real code deployment, we have these two options. So, so for developing any new code or doing any core coding related activities, we either use the IDE integration.

16:00 We install an extension in VS code. And we also use CLI whenever needed. So, using extensions, we can't do everything. We need to be familiar with both these ways. And in this video, we are going to cover all these three options, cloud desktop, integration, and CLI. SDK, it is something very advanced and it is not about like using cloud code feature. It is mainly about building something that can use cloud code at the backend. So it is a different text stack or different way of using.

17:00 Then this is not very famous. So, so now without any delay, let's get started with cloud desktop and do our first interaction with cloud code. So to get started with cloud, what we need to do is first of all, we need to visit official cloud website, which is cloud.ai. And here we can talk to cloud for free. Various cloud models are available here. And then we need to go here and click on upgrade plan. And if we click on this upgrade plan, we will see three options initially.

18:00 There is a free option, then there is Pro, and then there is Max. So currently I'm already using Pro. So it is not showing the free option for me. But from free, we need to go to Pro and Max. So Pro is the one we have to buy and use to use cloud code with. This is like the minimum plan that we need to use. So cloud basically they don't provide this cloud code is free. So there is no trial version or something.

18:40 So don't spend your time exploring like how we can use cloud code for free. Because they don't provide any free version. So to use that, we need to buy this Pro plan. So we shall go ahead and buy this, and then only we will be able to use it. And they have a monthly and annual plan as well. So you can go for monthly plan and at least for one month, I will say go ahead, buy it, and then do this exercise and try it out.

19:20 It is very well worth it spending our effort. So that you try out the real thing and then only you realize what is the power behind it. And once you have this Pro plan, then we can go back from here. And then we can go to this Get apps and extensions. So if we click here, we will see that there is entire ecosystem of cloud. It is not only about cloud code. Cloud code is just a part of entire cloud ecosystem.

6:20 So first let's understand what exactly is this ecosystem and what all comes with with this cloud and the Pro plan. So here it is an overview of the entire cloud ecosystem. So when we say cloud, not the cloud code. So cloud, it is an integrated AI platform for chat, chat, coding, creativity, productivity and whatever enterprise development that we want to do. And at the heart of it of everything, these are the foundation models, which are powering every cloud experience, whatever query we are running, whether we are asking it to do anything, whether we are chatting with it, whether we are asking it to develop code for us, whatever we are building, so So they cannot provide any unlimited plan.

5:40 So they have this plan with some limits. And currently we can see that the limit, it is 4%. And it will reset in 4 hours. So here we can see that limit as well. And here we can see some more information about the usage like currently whatever we are using, then what exactly is consuming our usage. So the agent browser. Like the Claude desktop that we used just now. So it opened in separate browser when it was testing the game.

6:20 So that is what it, it is like it contributed most. And then there are some more things that it did internally. So like that, we can check. We will understand these things more skills, plugins, MCPs and all, then you will understand like what exactly do they. So that is why creating this file, it is very important and it is, we use it in each and every project. And the structure, we already saw, it looks something like this where we have an overview, some standards, testing instructions, architecture, what things it shall not do and all.

6:20 So it is included in every chat that we will do with. And these are some of the key contents that shall be there, or as a best practice, it is followed. Like the our core project structure, it shall be there, what are the main folders and important files. What are the coding standards we follow like naming conventions, any pattern or something. How shall test our code, how it shall run it, like using virtual commands, some specific commands or any, let's say Docker kind of things if we have to mention how it shall build the project, if we have to standardize, what workflow it shall follow, or if some specific task we wanted to do.

7:20 If we are So it is not specific to Claude code. It is like the way of talking agents with each other. So even before Claude code, like when we have normal LLM and agents for that. So whenever they had to share something or extend their capabilities. So every company was trying that part in their own way. And that is where Anthropic, they suggested this way, which was like really well and it was very well designed. That everybody was like impressed by this.

6:20 Everybody started using this protocol. And that is where like it became very famous. And now each and every agent that is there, so every, every agent they support this MPC. So this is installed now. And it is saying like run reload plugins to activate. So I will do the same. So reload plugins. It has reloaded it. And now if I, again like run the command plugin. So we will see the same. And if we want something else, more also, we can try.

6:13 And if I press the right key to go to this installed, I think we can use arrow key and I see it is clickable as well, which is good. Let's now work on our second task, which is K N3. And with this, the expectation is that now we will be able to upload the resume and run it end to end. So let's try to do this again, like similar way. This time it will probably be more smooth because we have already done one round of, like testing and everything.

6:20 And this time we will use a particular command. And what is that command? It is feature dev. And if you have guessed it, so this command, it is coming as a command when we installed the plugin. So, like I was taking talking, right? So whenever we install any plugin, so it can be a server, it can be a skill or it can be a specific command. So because this is about feature development, they provide.