Plandex: RIP Cursor! NEW Agentic Coder! AI Software Engineer Automates Your ENTIRE Code (Opensource)

I think I just found a great alternative to cursor and it's not augment code. This time around we have a new AI coding assistant that is fully open- source and it is capable of handling large code bases. Allow me to introduce Pandex. Pandex is a terminal based AI coding agent which is purposely built for realworld software projects. It can autonomously plan, execute, and debug large coding tasks across dozens of files with multi-million token context support. Plandex is something that is also open- source, which is where you can use all of the functionalities completely for free locally. There's a lot of key features I want to highlight. Firstly, is its massive context window where it has a 2 million token context window with support for indexing over 20 million tokens using Treesitter. This is going to be allowing you to have AI make changes across your fullon code base. You have safe editing mode. You also have an autonomous and controlled mode, which is where you have a full auto mode or a step-by-step planning mode, which you can have the AI work with on your side. You also have multi-model support for seamless combined models from various providers like OpenAI, Enthropic, and many other open- source model providers. You have project where chat where it can understand your repo and help brainstorm version control. It's going to be something that you can easily set up within your terminal. And you can actually set this up with a oneline cl install or with docker. And this is something that we're going to be taking a look at throughout today's video. To showcase what plex is capable of doing, just take a look at this ability to autonomously debug browserbased applications. In this video, you'll see that it is capable of identifying and resolving issues in the codebase by planning, executing, and validating changes across multiple files. Regardless of how many files you have, you can have the AI deployed to scrape through all the file contents and then work on validating changes across them. In this case, you can see that Plx diff review sandbox which will allow developers to review AI generated changes before applying them will ensure code integrity and it will then work on making sure that the code is swiftly debugged and fully processed. Before we get started, I just want to mention that you should definitely go ahead and subscribe to the world of AI newsletter. I'm constantly posting different newsletters on a weekly basis. So this is where you can easily get up-to-date knowledge about what is happening in the AI space. So definitely go ahead and subscribe as this is completely for free. How it works is actually pretty interesting. This is a diagram that outlines the Planex AI development workflow which guides how it tackles large scale coding task step by step. It first starts off with the chat node. This is where developers can start brainstorming, asking questions, refining the task until it's well defined. From here, it moves on to the next node which is the highle planning context selection. Essentially, here is where Planex maps out the task broadly and it selects relevant projects with the context that it needs using its tree sitter map. From there, it goes over to the detailed planning step. This is where it breaks down the task into small manageable steps with specific file context. From there, we then move on to the implementation plan, which is where it's going to be doing this within an isolated sandbox. This is where the code changes are made inside a version controlled sandbox and it's going to be isolated from your main project. The fifth step results in the review and the revision process where developers can then inspect diffs in a local UI or a browser, rewind changes, reject mistakes or try alternative models. From there you can move on to the sixth step which is the tentative apply and debug process where changes are tested on the real project itself and if they actually fail rolls back to the previous uh viable option and from there it will assist in debugging and lastly we have the commit feature which is where once everything works the AI generates a commit message and finalizes the changes and this is essentially the loop that it works on to iterate and allows for flexible adjustments, debugging, multimodel testing, all while keeping your codebase clean and in control. Now, I know with other models as well as other agents from like something cursor has, it is not always properly going to be able to make changes across your overall codebase. It might make a lot of errors, which is why people are looking for other alternatives like Planex as well as augment code. Now, if you're wondering how to get started, there's three ways to do so. You have first the Planex cloud, which I believe is paid, but it is something that will easily get you started. There is no need to set up anything because you can easily just

access it off of the web through their own hosting. Now, another option is Plantex Cloud, where this is where you bring your own API key. You use your own uh API key. You still use the Plantex cloud service, but you use your own API key for the generations. And then lastly, you have the self-hosted local mode. This is where you can run Planex locally with Docker or host off of your own server and then you can even bring your own API key or use different providers like OpenAI or Enthropic. Now it is something that you can install for various operating systems whether that's Windows uh using WSL, Linux as well as Mac OS. Now what we're going to be doing is installing Planex with the self-hosting method. So, it is going to be completely free. And what we're going to be doing is installing it with Windows. So, we're going to need WSL if you're following through with the Windows operating system installation method. So, go ahead and install this. This is a subsystem for Linux. Once you have that installed, make sure you have Git installed as well as VS Code. Once that is done, go over to the doc, which I'll leave a link to in the description below. Open up WSL. And then once you have done so, you want to go ahead and clone the Planex repository. So, copy this link and then paste it into the terminal. And once it has finished cloning, you can then go into the Planex folder as well as the app folder. For this next step, you'll also need Docker Desktop installed. Make sure you install this for your operating system. In this case, install for Windows. I have it already installed, but what you're going to need to do is simply go ahead and open it up. So, open up Docker Desktop and make sure it is running in the background. And once you do have it running, make sure you go back into the docs and you want to go into the main Planex directory. And once you are in there, you want to copy this next command, which is to start the local server. And once you have started it, we're going to need to go enter in a new or open up a new uh command prompt. So, open up another WSL terminal. And once it has finished starting up the server, we can then go ahead and run the next command which is to install the Planex Cly. So we have now locally started up the Planex server and you can see that it is running on our 8099 port. But what we're going to need to do now is install this Clive with this command. So use this curl command and paste it into the new terminal that we have opened up. And you can see that it is going to ask you for your password. Once you have inputed your password, it will then install this. Next, you need to go and copy the Planex signin command. And you want to go into a new terminal and paste this in and click enter. Now, you don't want to select the cloud version. You need to select the local mode host. And you'll see that there's a local host already highlighted, which is a server that we have already started up. This is the local host uh 8099 that is available. So what you can do is simply just paste in this local host and click enter. After you have created an account with Planex and it is fully locally hosted, what you can do is set up your API key. Now you can use Olama to use a local model. You can use other providers like Open Router or OpenAI. You just need to simply use the expert command to set the API key for the provider you want to use. And once done, you can just use the Plantex command within the terminal to start it up. I've now set my key and I'm actually requesting it to generate a SAS landing page. So let's see what it actually does. But what you can do after you have configured your API key and you have provided the actual command to start it up which is the plex command. You can actually set where you want to deploy Pandex into. You can configure Planex so that it indexes all the files and it can then make changes across the fullon directory that you have added. So this is where you would want to use the slash command to see all the commands. You can also use your docs as to how you can set this up. But in this case, I'm creating a new directory to create this last landing page. And once it has finished doing this, I'm going to go ahead and open it up. So right now it is actually generating the plan for it. and it will also display all the different files it's going to edit. So in this case I have it within a bolt. diy uh folder and I'm going to be creating the app within that directory. So you can see all the different files that are loaded into the context. And this is why this is a great tool cuz it's able to index all the files that is necessary for this and it will then execute the commands based off the plan it creates within your terminal. And there we go. This is the SAS landing page that it was capable of generating. It has animations and it actually looks really, really good. It's an intuitive SAS landing page. You can see there's a testimonial roller, the carousel that is actually animated. And it hasn't finished generating the rest like the flexible pricing as well as the other components such as the footer. But overall, it did

a good job in generating I guess the base structure of this app. So, I really love the capability of this uh Planex agent and it's something that I highly recommend that you use. This is just one example and this is not even the main purpose of this uh tool. It's mostly for uh implementing large changes across code bases. So, this is something I highly recommend that you take a look at with the links in the description below. Make sure you take a look at their GitHub repo. Uh this is where you can learn a lot more about how you can set this up. Follow me on the newsletter, join our private Discord, follow me on Twitter, and lastly, make sure you guys subscribe, turn notification bell, like this video, and please take a look at our previous videos so that you can stay up to date with the latest AI news. With that thought, guys, have an amazing day, spread positivity, and I'll see you guys really shortly. He suffers.