Hey everyone, what's going on? Welcome back to Code with Nathan. I'm your host, Nathan, and I simplify complex tech topics so you can master them easily. In this video, I want to introduce you to a cursor alternative that's open- source, local, and most importantly, free. Now, if you watch this video, I'm pretty sure you are already familiar with cursor, but just in case you didn't, cursor is a premium code editor with AI feature. It enables you to build application by making use of AI models to generate files, write code, refactor existing code, and debugging errors. Now, many people love Cursor for its AI powered features, but the free version can feel limiting, and some people are just not ready to commit $20 per month for the pro version. What's more, Cursor itself is actually a fork of Visual Studio Code, or VS Code for short. VS Code is a code editor developed by Microsoft and loved by professional software developers including me because of its simplicity and extendability. It's one of the best code editor applications to this day and because it's open source. Many people contribute to the project to improve it and make it what it is today. Some people feel that cursor just took VS Code, add some AI features on top of it, then sell it for profit. There's nothing original or great about the editor. Now, personally, I'm not a cursor hater. They don't violate VS Code license agreement, so they're free to do what they want. But I get that some people just don't want to pay for it yet. Maybe they're only using it for hobby projects, but the free version feels too limited. So, for those of you who would like an alternative to Cursor and save that $20 for something more important, the solution is quite simple. You just need to install VS Code as your code editor, then integrate a couple of extensions to get the same features as cursor. Essentially, there are two extensions you need to install, which is client and continue.deaf. This setup is even better than cursor because you have full control over your data. You can use free and local models and keep costs as low as zero. Plus, you don't have to worry about enabling privacy mode to prevent cursor and third party companies from using your data to train their AI models. I'm going to show you exactly how to set it up. So, please watch the video until the end. All right, let's get started. First of all, if you haven't already, you need to install VS Code on your computer. Visit the VS Code homepage at code.visisualstudio.com visualstudio.com and you will see a button to download the version for your operating system. In my case, it showed the Mac version. Download and install it and then we'll move on to the next step. Next, you need to install Olama to access AI models locally. You can head over to.com to get the application. I will put all relevant links in the description as well. With Olama, you can download state-of-the-art open- source models locally, such as the DeepSc R1 and Quen, and install them. This way, you don't need to have an internet connection every time you want to access the models. Now, keep in mind that using models locally means you need to have enough storage to keep those models. If you have limited memory or slow internet access, then you can use the Open Router API as an alternative. I will show you how to use open router later after we finish the setup. Now that we have VS Code and Olama installed, open VS Code and click on the extensions tab right here on the left side. First install continue deaf. Just type continue DAV in the search bar to find it. Then click the install button. Once installed, search for client and install it as well. These two main extensions will essentially replicate the features of cursor and VS code. Now that our extensions and code editor are installed, we can configure and add our models. Open the continue extension. Close the create assistant model as you don't need it. And then click the cube icon over here to configure the models used by continue. The models for continue will be provided by Olama. For the chat mode, you can use the basic Deepseek R1 model with 7 billion parameters. This model is quite good in answering coding questions. For code completion feature, download the Quen 2.5 coder with 1.5 billion parameters. A small model is enough for the autocomplete feature as larger models have been shown to not significantly improve performance. To download the models, you need to open the terminal or command prompt for Windows. Then run the command Olama pull followed by the name of the model you want to download. In this case, it's DeepSync R1 and the next one is Quen 2.5 coder. You can pause the video here and take a moment to download the models. Once you downloaded the models with Lama, head back to continue and put the model you have in the models parameter. Put the name, the provider, and the model for deepse R1. Save the changes and the model configuration should be pointed to deepse. Next, do the same for quen 2.5 coder. You also need to add the role for the quen model as shown here. Save the changes again and now the chat and autocomplete feature should be working. Let's test the extension by creating a simple Python file. I will save this file in my desktop and name it as sample.py. Create a function to add two numbers A and B. And there you go. You can see that the autocomplete feature from continue is working. Press the tap button to accept the suggestions. Now let's test the chat mode to see if it works. I will instruct Continue to create a Flappy Bird game. Here you can see that instead of editing the files directly, continue chose to respond to my request like a chatbot with no editing capabilities. Let me just cancel the request here. One important thing to mention is keyboard shortcuts. If you click on the Jer icon and then shortcuts, you'll find helpful comments for highlighting code, getting suggestions, and autocomp completion. While continue works well for code completion, it certain features such as multifile editing and running commands from the terminal which is available in cursor. By multifile editing I mean that I can just give it a prompt and it can go through all existing files in the currently opened folder create necessary files and complete the application or promp I want without the barrier of which file I have currently open. This is why we also need client to fully copy the features of cursor. Client allows the AI to reason and think better before running an instruction. It can also handle multiple files editing better when compared to continue. Next, click the client icon. Here you can register an account with client and get trial access to powerful models. Just sign up for an account using either Google or GitHub. Authorize access to Klein and now you can start using Klein with free credits. All right. Now that we have everything properly set up, the extensions, the models, and the configurations, you can start using VS Code as a powerful alternative to cursor. Client can create necessary files, write your code in that file, and install dependencies just like cursor's composer. Continue.dev enables auto code completion, multi-line edits, and over features similar to cursor's prompt bar feature. Using continue, you can write some code, then highlight that code and ask continue to update it. It will generate the code for you, allowing you to insert or copy directly into your file. Let me demonstrate the capabilities of these tools within VS Code. In this case, I will instruct Klein to create a Flappy Bird game in retro style. It will use HTML, CSS, and JavaScript. The AI model will take a moment to process the request, generate a step-by-step plan, and create the necessary files. You can also enable auto approve to let client automatically run necessary actions such as creating files and running safe commands as shown here. Once done, don't forget to click the auto approve checkbox. client will take a while to generate the code and test the output by itself. So I will skip ahead to when the request is completed. All right, now my request to create a Flappy Bird game has consumed all of the free credits that I get from client. Let's see if the game can be played. I will open the HTML file in the browser. As you can see here, the Flippy Bird game is already playable, although it can be improved further. This entire project was developed with just a single prompt. So you can see now why client and continue can be used as cursor alternatives that you can use for free. If you want to utilize cursor's controllable K feature, you can use continue dev's alternative shortcut which is control I or command I for Mac. You can then ask continue to edit just the highlighted code. It also includes shortcuts for debugging, cancelling requests, and editing highlighted code. When your trial credits from client is exhausted, you can use free models from open router. Let me show you how to do that next. Open router is a web platform that provides a unified API to access a wide range of AI models from any providers. With open router, you can access many AI models using just a single API key. There are free and paid models provided in Open Router. And to use the paid one, you need to buy some credits just like all the other model providers. We're going to use just the free models offered by Open Router. So, first you need to sign up for an account. Then, create an API key. Next, copy the API key. Then get back to client and change the provider API to open router. Paste the key and select the model you want to use. It's recommended to use a large model such as the Deepseek R1. This model is the same as the one from Olama. Only this one has the 671 billion parameters, making it capable of running prompts given by client. And that's it. Now you can use client with dipssec as the AI model. You can also check out other free models offered by open router such as dipssec v324 which has 685 billion parameters. This dips model is very powerful and seems to be on par with GPT4 only in code execution and refactoring. Now, keep in mind that because these models are free, you may experience a long request time or even a timeout when many other people also use it. So, please be patient if that happens. And now we have reached the end of the tutorial. So, what do you think? Do you prefer to use VS Code with client and continue defaf or do you prefer to use cursor? Let me know your thoughts in the comments. I will join the conversation and reply as often as I can. If you like this video, consider to like the video and subscribe to the channel. I will be creating more videos on AI coding and tech related stuff as soon as possible. So, make sure you turn on the notification bell so you won't miss any new videos. With that said, have an amazing day, keep learning, and I'll see you guys again soon. Thanks. [Music]