NEW Warp 2.0 AI Coding ADE Beats Cursor & Windsurf? CLI AI Agent (Workflows, MCPs, Prompts, & More)

What's going on guys? There has been a lot of new updates lately in the AI coding space. We just had Google Gemini release their CLI tool. Of course, we've had Claude Code. We have OpenAI codecs. Many different tools. Then we of course have our IDE, Cursor, Windsurf, a bunch of different tools out there. And there's a lot of different ways people are going about their coding experience. Now, in today's video, we're going to be taking a look at kind of like

A little new hybrid approach for AI coding. Now, this approach is what they like to call the agentic development environment. So, ADE. So, instead of IDE, it is ADE. And this tool that we're going to be covering in today's video is called Warp. They just launched their 2.0 agentic development environment. So, we're going to take a look at it, see what it's actually made out of, and see if this could maybe be a new potential way to actually code with AI.

Let's dive right in. So, a lot of you guys may be wondering, what is an ADE? And we're going to dive into that in just a second. But before I do, just a shout out to you guys who actually commented on my video yesterday on Open Code mentioning Warp. Okay, you guys mentioned that Warp, a lot of you guys were saying it's pretty good. You like it. Some of you were saying it's kind of bloated. We're going to test it out

And see what it's really made of. Now, this is Warp's website. All links I cover in today's video will be linked down below as always. And as we can see here, the Agentic development environment. Warp is the fastest way to build with multiple AI agents from writing code to shipping it. the best overall coding and terminal agent. So, it's very simple to actually get started. You can download it for Windows right here or run Windet install warp.warp and you can see

Here trusted by over half a million engineers at some of the leading companies and they have a new 2.0 launch video right here from the founder and CEO. If you're really looking into this, I definitely recommend checking it out. Given some kind of features and whatnot here, they have a nice looking website that always helps. and number one on terminal bench and swb bench verified 71%. So if we look at terminal bench here and this is the terminal bench leaderboard uh

We can see that warp is the number one agent with multiple different models. So you can see the date here and model oranthropic accuracy right here. And you can see number two is claude code with opus 4 and claude code with opus uh with sonnet 4 is number three claude code with sonet 3.7. So you can see all the different ones right here. Um we got codeex and I'm assuming Gemini will probably be on here anytime soon. Okay, there's actually some

Interesting ones here. Terminus, uh, and then Goose. So, we may check those out in future videos. If you want me to check those out, let me know in the comments down below. Also, let me know if there are some other ones that you want to see me check out because who knows, like you guys saw, I may just do a video on it. We get into Warp. The one I guess downside is that there's paid plans. Okay, it's not that big

Of a deal. Of course, many different tools. Cursor, Windsurf, there's paid plans, right? Um, so that's fine. If we're looking at their paid plans here, they do have a free tier, which I'm currently on. And this gives you uh up to 150 AI requests per month, which if you're going to be using this on a day-to-day is quite low. And I'm not going to read through this entire stuff right here. You can check it out on their website, but their pro

Plan is about 15 bucks per month. Okay, so it's actually 18 bucks if you're paying on the monthly version and then 40 uh 50 bucks on the turbo version. So this is of course if you have a large team. Um so yeah, there are I think unlimited next command suggestions. So, that's good. Three index code bases on the free plan, up to 5,000 files codebase. You get access to all the top models and a few other stuff here. So, it's not

A bad free plan if you're testing this out. Uh, if we look at that compared to Cursor, I guess it's two bucks cheaper than their uh plan right here for But Cursor, you get unlimited agent requests, unlimited tab completion, all that good stuff. Um, and then here you're getting only 2,500 AI requests. So, not bad. As a side note, I was also looking into the Wave terminal here, and maybe I'll do a video on it in the future. It's pretty good,

Too. Now, this one's open source. I think they have a lot of potential if they just add some of the They have like Wave AI right here, but I think if they add some of the features that Warp has kind of like that, which you're going to see in just a second. I think it would be really, really cool because I love this uh being open source. Don't we all just love open source tools? But anyways, let's dive into Warp. So,

Once you download it and you open Warp Up, it is going to look something like this. So, you're going to have to log in first um to Warp. So, you're going to have to create an account. It's going to ask you and prompt you to do all that. Then, once you do that, you're going to be in a terminal like this. So, an ADE, what they call it, is essentially like, you know, your terminal but with more coding capabilities. And I

Really do like the way Warp kind of set this up. I do like the UI for example. I also like this too. Like we can click this and we can easily switch to folders. So let's go to like development. This is like on my computer. Now we're in our development folder. That's really useful. Like I think that's really cool. Just have that UI there instead of having to do CD which I do all the time. But you can click the UI

And then you see that it runs CD development. Now there's a few different modes down here. So we have terminal. So this is if you just want to run straight terminal commands like your typical terminal. We have agent mode. So this is, you know, we're going to get dive into this in just a second when we start AI coding. But you can have fix, explain, install, code, deploy, bunch of different options. And then auto detection. Okay. Um, another cool thing too

That I actually really like is you have voice input. So this is good for you vibe coders out there which we all pretty much are doing some level of I guess vibe coding um depending on what your definition of vibe coding is and then only usable in git repository. So this would be adding stuff once you have a git repository and then attaching images. We have our different models up here. So, uh, Claude Sonnet 4 Light Claude, uh, Sonnet 3.7, 3.5,

GPT40, 03, uh, 04 mini, Gemini 2.5 flash. So, all the different ones that you would really want to use. So it's kind of set up like cursor where you have that certain level of you know tokens or usages but I guess more like wind surf maybe but it's not set up like um something where you can actually add in your own API keys which I think that is you know I like doing having that capability as well. That's one thing I

Do like with Wave. It actually allows you to add in your own API keys which is really nice. Uh of course that's how it would work with it being open source. So now that we're in development, I'm just going to CD into my sandbox here. And instead of using shell, right? So this is auto detection. So auto detection is kind of like a mix between agent and terminal. It's going to detect if you're, you know, using either one. So I could

I just used a shell command here, but now I can in auto detect say so I'm going to say make a new folder called warp-ext. And now we can see it's thinking it's using the agent model from that auto detect. And now it's saying I'll create a new folder called warp test for you. Uh okay. If I run this command and read the output. So we can actually say going to say make dur um warp test. You could either say cancel

Refine edit. Uh we're just going to go ahead and click run. Okay. The folder has been successfully created. Now let's change dur into it. Okay. And now you do have to be careful a bit cuz sometimes I guess the auto detection won't always detect like this. It's thinking that I'm trying to do a shell command when I was actually trying to do like a plain text agent command. Okay. So actually if I changed it to a little bit more of a

Conversational tone like here change the directory please into it. Now it's picking up that it is an agent command. So it is working that way. Now we can go ahead and run this. Now, we may not want to always approve things. So, we can actually go ahead into our warp settings. So, if I go to settings up here and go to AI, this is where you can change some of the AI commands. So, you can see active AI, some of this

Stuff. If we scroll down to permissions, so you can see your apply code diffs, your read files, your create plans, and execute command. So, I'm just going to change this instead of agent decides to always allow. Um, and then you have some command deny list. So different commands that you don't want it to do such as maybe like remove certain files or whatever the case may be. You can see also it has MCP support. So call MCP servers. You can add

An MCP server like so. Click on add. Here's your MCP server JSON right here. You can go ahead and do so. I'm not going to get into all the MCP stuff in today's video. So if we click on code here, we can see codebase index right here. Index new folders by default. You could toggle that if you want. Now if you go to teams, you can actually create a team, which is good if you're running a team. And you can invite

Your team members and appearance. You can do some of the theme, change typical appearance stuff, features. So you have some general features here, different session features, editor features right here, um terminal notifications, workflows, your keyboard shortcuts. Now if we go ahead here and click on this warp drive on the top left, um this is what they call the warp drive. So you can up here it will show create a team. I already created a team. So we have executive stride right

Here. I could go ahead and add uh folders, add workflows, add prompts, add notebooks, environment variables, import. Um, and you can add your team members here, which is really cool. So, if I create a folder, let's say prompts. Boom. Now, we have our folder here. I go ahead create new workflow. And you can see here, you could add a title. And then there's a command right here. So, you could add a description, a command. So like echo hello name where we're

Describing what this command is. Enter a single line command or entire shell script. So you can kind of build out workflows here. Add aliases right here. Add environment variables. So I won't get too deep into this right now, but you do have some example ones down here. So if I go um this is your person. So this is your team right here. And then this is your personal side. So you got MCP servers right here. So you can add MCP servers.

You got rules. So this is exactly how like a rules section would work in something like cursor. You can add your rules. So like force unwrap, always unwrap in rust or whatever. Like if you're using TypeScript, maybe you want to say use types over interfaces or whatever like typical rules that you want for your codebase, you put them in here. And then we have a starter workflow folder example. So we can see some example workflows. So if I click on this,

So if I click on this right here, we can see our example workflow um is right here. And if I click here, squash the last n commits together. So we can see this is the example workflow. You could edit the workflow. Squash the last n commits together. Squashes the last n commits together. Requires rewriting a commit message for the new squash changes right here. Um, and then yeah, so you could like get pretty I guess advanced or do some cool stuff

With these workflows and I think they're pretty cool. Um, and then notebooks. These are essentially like Jupiter notebooks that you could have right here and you could follow along. So, uh, run in terminal. Um, so you could document certain processes maybe for your team or just internally for yourself. um maybe like let's say uh installing something or running something, hosting something, self-hosting, whatever the case is. And boom, getting started with notebook. So really cool stuff. I do like this sort of

Interface. Personally me, I love still like I love a lot of things about cursor and the AI IDE space like and be able to see the actual files, but I do like a lot of stuff with this. I personally like sort of like a hybrid of um different tools. Um there's pros and cons to each. So that's my uh opinion. You know, I use maybe certain tools together if I find that there's enough value with different sides of things, right? Maybe

Using clog code with cursor, using warp with cursor, right? So there's a lot of different options. Let me know what your typical coding tech stack is in the comments down below. What are some of the tools, maybe low-key ones that maybe aren't out there yet that you may know that you really like or ones that um that you just kind of like using together and you think other people could benefit from that stack? Let let me know in the comments down

Below and maybe we can get some uh talks going down there. Okay guys, so now we covered some of the warp basics. Okay, we went over pretty much a lot of the different stuff in terms of how to use warp and kind of just how it all works together. But now, let's actually try to build out a basic project. All right, so here I'm saying I want to build an X.js app that is a fitness tracker. It should track calories and

Have some pretty cool features. No backend at the moment, just the MVP to start. And we're going to go ahead and click on send. Now, you could attach context if you're, you know, doing certain things. Like if I click attach context, we could um add this like this command attach as agent mode context or any other command at the top which is pretty cool that you can do that. And you can also save commands right here as a workflow or filter

Block output. And here you can do things like copy command, share session, um a bunch of different stuff here. So a lot of different features that really put your terminal on steroids. Okay. So, we can see here that we are viewing a command details and I'm going to go ahead and allow this. So, if you click control shift I, it will auto approve all agent actions for this task. I'm going to do so. Okay. So, we're initializing the project with the

Template here. Okay. So, index codebase. I'm just going to click on index codebase here. Now we're installing some dependencies and it's creating the main application structure. Starting off with the main page. Okay. So it created the component directory here and now it's creating the calorie tracker component. Okay. And we can see all the diffs right here which is actually really nice. I do like the interface. I do like how it's collapsible. So it's not just like spamming you with diffs like

Some uh other tools may. I'm actually really enjoying the experience though within this. like the user experience. It's to be honest, it's a lot nicer than a lot of times when you're coding within the IDE like you know with some other tools. I really like a lot of these other tools like Ader, Gemini CLI, Claude Code, but uh sometimes it can get a little messy I find. So this user interface is quite nicer. Wow. Okay. Boom. So we got a bunch

Of stuff done here. um worker logger component with exercise tracking progress dashboard with charts and analytics. Uh fixing syntax errors in main page fixing TypeScript errors um and then boom, we got it to start the development server. So let's take a look at the first launch of this and see how it actually is. Boom, we got our fitness tracker. So actually really good job out of box here. Getting a hydration error right here, but that's fine. uh calories, workouts. I actually

Did a really good job for a first prompt. Like that was literally one shot prompt here. You can see total workouts, hours exercise, calories burned, all that good stuff. Calories here. Um and then workouts. So we got like baseball, boxing, rowing, etc. Okay. So another feature I want to show you is the parallel agents feature. So you can have multiple threads with agents working simultaneously. And this is really cool because you know some tools won't really allow you to do this.

And so let's go ahead on this tab. And all we're simply doing is just having two different tabs here. So I'm going to say I want to make this way nicer, way more modern, and have dark mode and light mode toggle. And then in this one I'm saying create a calendar tab too and make there be a sidebar instead of the menu items on the top and make it look nicer. So, I'm just going to go ahead and click send on

Both of these. And then we're going to have both of these agents working at the same time. So, you can see here there is the uh purple or the pink icon right here, meaning these agents are working. We go ahead and click on agents up here. You we can see what's actually going on with these agents. So, we have one creating the calendar tab and then the other working on the light mode and dark mode. And if there is input that

Is needed from us, it will actually let us know what agent needs our input. Of course, this is pretty powerful because imagine having a bunch of different agents, a bunch of different team members working all at once. Now, you probably want to do this at a small scale, especially first starting out, so you're not just completely making all these changes to your codebase, but uh as models get more and more uh powerful, I mean, think about it. You could have a

Full development team working uh on your codebase. All right, so we can see both of these agents are working pretty hard right now. Uh going through many different files and making some changes. All right, so now looks like we got a bit of an issue here. The whole screen's white. So I basically just said, uh now the whole screen's white and don't start the development server. I already have it started. So it's seeing the issue. Um it is trying to fix

It. That's one thing you may have to be a little careful if you're running multiple agents in parallel because they could be changing different things and some agent could change something that another agent has changed, maybe broke your app. So, I would say be a little careful with this. Maybe do two separate tasks that are very different and have those agents work on that. Maybe front end, back end, or whatever the case may be or different features specifically. Okay. And boom.

It's saying that it successfully fixed this. So, let's go ahead and see. Okay. And boom. I'm actually very impressed. I'm not even kidding. I'm actually very impressed by how good it did. And like the animations right here. Okay. The light mode and dark mode isn't working. It just has a toggle here. So, that's one issue. We can see the sidebar looks pretty good. Got some issues here. Calendar looks pretty good though. We can click on these. See the calories. Of course,

This is all within a couple prompts. Um, we just had our multi- aents work on this and whatnot. So, all in all, I'm pretty impressed with Warp. Um, am I going to be using this? I mean, I will probably be using a bit more, testing it out. I like these terminals, and I will try out some different other terminals on this channel as well. So, make sure to stay tuned for that. I do think, you know, a lot of these different

Tools like ADE, IDE, Terminal CLI, whatever, they have their pros and cons. And me personally, I like using a combination of some of the top different ones. I think a hybrid approach is pretty cool and you can use it in different scenarios. But let me know, like I said, what your thoughts are in the comments down below about this. Have you tried Warp? Do you like it? Do you not like it? Do you like something else better? Or do you like

Just a plain old simple ide? All in all guys, that's pretty much it for this video. I just wanted to kind of dive into this tool. I know a lot of you guys were recommending it from my last video. And if you have any other ideas, let me know in the comments down below. Guys, check out some of the links down below in the description if you want help with growing your business with AI or check out some of the stuff

That we do at Executive Stride. But other than that, I will see you in the next video, guys. Keep hustling, keep grinding, and of course, guys, accelerate your stride.