How I Learn Claude AI on Anthropic Academy | Free Claude AI Courses 2026

0:00 Hi. During this tutorial, me and you together, we'll learn Cloud Code following Anthropic's official course. If we haven't met, my name is Annie. I'm a tech consultant based in Stockholm. Let's get started. If you want to follow along, you can get the same link in the description. So, right now, I'm logged in into Anthropic's Academy, and from there, I choose Cloud Code in Action course. And to start with, we are looking at what is a coding assistant. So, to follow this documentation, when it comes to large language models, those models can only take text input and then return text output. So, the question is, what is happening when we ask large language models like Opus, Sonnet, or Haiku to read information from a document?

0:50 So, at that stage is coming along another tool that is called coding assistant. It gathers information on the context, formulates a plan, and then it takes an action. The example will be when you send a question to Cloud, you ask about something that is saved in a document. And then the code assistant will send an instruction to Cloud and will ask Cloud to ask information about this document. For example, if you ask Cloud, what code is written in the main.go file, then the code assistant will add underneath this input, if you want to read a file, respond with read file, name of the file.

1:37 So, code assistant will help Cloud to ask for the correct tools. So, the idea here is that every large language model can take input in a form of text and then can apply a lot of knowledge on this text and answer with a text output. However, every time you want to use another tool, for example, you want to write something, you want to read something, then you need an assistant from Cloud coding assistant. And this is how the coding assistant works. So, the benefits of Cloud coding assistant is that you can use a lot of complex tool, and which means that you can complete a lot of complex work. Then, last but not least, it is quite safe using coding assistant because while using coding assistant, you do not need to do code indexing.

2:28 For example, when you use GitHub workspace, in there, normally all the code in your GitHub is indexed, and different chunks of information are saved normally by the supplier on supplier's side, and this is not very safe because all the information is saved on the supplier's side. But when you use Cloud code, Cloud code inside your computer will use general commands, for example, find to get the right information, and only the correct information will be sent to vendor where Cloud API is hosted instead of just sending all the code base. So, that is another reason why coding assistant in Cloud code is proving better security. But in the next step, we will learn about Cloud code setup.

3:15 In order to set up Cloud code, you can visit this URL. I will save the URL in the description below so that everyone can access it. And once you go to the URL, then you will use the command that is used for your operating system to install Cloud code. For example, if you are using Mac OS or Linux or WSL, then you can use this command. If you are using Windows PowerShell, then there is another command that you can use. And if you are using Windows command prompt, then you will use this curl command. And then there is also another option for Mac OS. And after installation, you will just run Cloud at your terminal. And the first time you run this command, you will be prompted to pick a color for the terminal and authenticate with your Cloud AI credentials.

4:03 If you get an error that Cloud isn't found after installation, you hit a network or permission errors. And for that reason, you will look into troubleshoot installation issue in the documents. In the next step, we will look into adding context. So, in here you can see that Cloud is showing that for accomplishing a certain task, Cloud needs some information. However, it can be so that in the same folder where you initiated Cloud code, there are a lot of different documents. And Cloud doesn't need to have access to all of those documents. For that reason, it is very important to give Cloud correct context. And based on Cloud recommendation, every time when you initiate your code in a project where there exist code, start with init command.

4:53 And what init command does is reading all the code in your code base and creating a Cloud MD file that will summarize information in your project. Then there are different types of Cloud MD files. There is Cloud MD file that is created after you run init command. And it will have information only about the files that are in the project. Then there is another Cloud file that is called Cloud.local.md. And this is not shared with other engineers. It contains personal instructions and customization for Cloud. So, in here, you can enter information that you do not want to share with other engineers and is very specific to you. And then there is another Cloud file that is used with all projects on your machine.

5:42 It contains instructions that you want Cloud to follow on all projects. So, you can customize how Cloud behaves by adding instructions to your Cloud MD file. For example, if Cloud is adding too many comments to code, you can address this by updating the file. You can write edit Cloud MD directly in your editor or run {forward slash} memory inside Cloud code to open the file and then add an instruction like use comment sparingly, only comment complex code. The next thing that I want to tell you about is tagging the files. When you need Cloud to look at specific file, use {snobble a symbol} followed by the file path. In the next section, Cloud is talking about mentioning or tagging a certain file.

6:32 So, if you already know in which file is the information, then you can just tag the file. For example, you would say, "How does the auth system work?" And then you will tag the file where you know the information exists. And in this way you are referencing the file and it is easier for Cloud code to get the information because it doesn't have to rotate in all the files in the projects to find the correct information.