Emergent Labs: Build production-ready apps through conversation

--- type: transcript date: 2025-10-15 youtube_id: CJiXoZGnmQk --- # Transcript: Emergent Labs: Build production-ready apps through conversation In this video, I'm going to be showing you Emergent Labs, which is a platform that allows you to build and deploy full stack applications completely with natural language. In the video, I'm going to go through a demonstration of how it works and then I'll also walk through some of the features of what you can do with it. I'm going to say I want to build out project management software and specifically what I want to be able to do is have a canban view and I also want to have the list view. I want to be able to invite different colleagues and specifically I want to have the functionality that would typically be included in something like Asana or ClickUp or those types of platforms. Now, before I send this in, I just want to go through some housekeeping of the platform. You will be able to add things like different design or PRDs as attachments. You will also be able to connect your GitHub account if you do want to sync this with GitHub. From there, you're going to be able to have the option to choose whether you have a public generation or private. Now, additionally, what you're going to be able to do within here is you are going to be able to add your own custom MCP servers or additionally incorporate things like notion, superbase, or include memory in the process of building out your applications. As soon as we send that in, the first thing that it's going to do is it's going to provision all of the different cloud resources that is required for our environment. It will go ahead and scale up everything that we need. And then from there, the first thing that the agent will do is it will ask some clarifying questions within here. The agent is asking some questions. Please confirm it further to continue. So we have the core functionality priority. Which features should you focus on first for the MVP? Task creation, editing, status management, so on and so forth. Within this, I'm going to say I want to focus on the task creation, editing, as well as status management to start. I want to make sure to have user authentication. And we don't need to actually worry about the team collaboration features just quite yet. That is going to be something that we address a little later on. And then finally for the task management feature, I want to include a basic title, description, status, and assenee. And don't worry about things like advanced features like due dates quite yet. Let's just focus on the overall design and this initial capability of the platform. Now, as soon as I send that in, now it will take a little bit of time to process everything that you asked for. Obviously, depending on both the complexity of how much you've actually sent in as a part of your instructions paired with the actual complexity of what you're asking for in the application, it might just take a few minutes. The first generation, what you are going to be able to do is you can watch through the process if you'd like. Alternatively, you can just leave this open in another tab and it will go through and autonomously start to build out your application. What we see here is the agent beginning to understand the instructions. Additionally, within here, you can see that it is going to write out the back end of our application within Python. We have the server.py and then we also have the structure of the front end that is going to begin to scaffold out for us. So then from there, what it's going to do is it's going to think through and iterate through all of the different coding pieces that it needs to set up on both the front end and the back end. As soon as that's done, I'm going to circle back and we'll take a look at the initial generation. Next up, a really neat feature of the platform is they do also have a testing agent where what it will do is once it has a understanding of your codebase, it will begin to go through the process of testing out the backend endpoints all the way through to actually testing out the front-end portion of your application. Here we can see that there's browser automation going to the login page of our application. And what's really neat with this is it can actually test to make sure the functionality is in fact working. And in this case, you can see that it's actually making a test user account with the email and password completely from scratch. And we can see that once it's through the process that there is in fact a functional canban board. We have the list view as well. And all in all, we have a functional application being tested all autonomously with AI agents. This is a really neat process. it will recognize that and it will go through and effectively pass those errors into the agent to further iterate on whatever needs to occur. It basically end to end is able to take everything from the PRD or in other words the instructions of what you have for the application go through build out the backend the front end and also do a critical part that you don't see too much of these types of platforms do is actually perform a very thorough autonomous QA process which is definitely a very cool feature to see. Now additionally what you're going to be able to do within this is you will be able to share this. If you do want to share the preview link you're going to be able to do that. Now, additionally, you can open this up in a new tab, which for full stack development obviously is something that is quite helpful. Now, one of the cool things with the platform is you will also be able to create mobile applications if you'd like. That is a part of their paid tier. So, that is something to be mindful of. I'm going to go ahead and let this run through the process and then as soon as it's done, I'll circle back once it has finalized everything and we'll take a closer look at our application. In between actually editing the different cuts, the agent has actually been running autonomously for about 15 minutes. And that's not to say that the agent is slow. This is actually one of the most thorough agentic experiences for building out a web application that I've seen. If I take a look again, it went through, it fixed those initial bugs, and then from there, what it's doing is it's actually going through and doing a further test. We can see that it's actually adding things to our canban board. If I progress through the process, I I can see it's toggling over to the list view. We can see that the list view is all functional here. All in all, I am super impressed with this platform. Now, as you might imagine, you are going to be able to continually iterate on this with your vision. And as you might expect, it will go through the process of actually testing everything out and making sure that everything is functional. And now the one really great thing with this that I actually quite like in having built these code generation tools before myself is one thing that you really have to be mindful of is that a Gentic process after the LLM has generated all of that code because you basically are just hoping that it has generated what you wanted. But with this, it's really taking it to the nth degree of what is capable right now because you're going to be able to effectively pass this off almost as if you're handing it off to a QA engineer to go through and do all of their own individual checks and then cycle that back to the developer team if there are any edits that need to occur. And the process will go back and forth until those conditions of the PRD or what you had asked for is actually met. It is very thoughtful in my opinion in terms of how this was actually built. This does definitely replicate a lot of the motions of what a development team would do. So overall, I definitely really encourage you to check out the platform if the testing agent is still running. One thing that I do want to demonstrate is you will actually be able to test out your application. So within here, I went ahead and I made a new account. And within here, you can update the status of all of the respective tasks. We can see that the edit functionality is also all set up as well. And if I want to add in a new test and I say hello world and I add in 1 2 3 for instance and I add that to my to-do list I can see that it is all functioning. And then just to test this out just to make sure that the persistence of the application is in fact there. If I log out of the account and then I go ahead and log back into the account I do see that the card that I created is in fact all there. It is something that if I were to have to spin up a prototype and actually trust the process of what's being generated. This is definitely something that I would strongly consider because just knowing the fact that there is an agent that will be spun off and do all the testing without effing to actually worry about it does give me a tremendous amount more confidence in the platform. Next up, in terms of some other features of the platform, you will be able to have multiple projects running in parallel and they do have this tab feature. For instance, if you want fully functional applications or build multiple applications in parallel, you will be able to do that. You can see this Twitter clone that I created shortly before demoing this application for you and you can just go and you can spin off a new idea by making a new tab effectively. So, it is very ergonomic in terms of navigating the platform. And additionally, in terms of some of the other features, if you are interested in trying this out, it's 20 bucks a month for their standard plan. They have also some more premium plans as well. If you do want to have things like more generations as well as accessing some of the state-of-the-art models out there, you will be able to access that for $200 a month. Their standard plan is very similar to some of the competitors that are out there at $20 a month where you'll be able to get a 100 credits per month. And just to give you a sense in terms of how many credits this does actually cost. So for this application that you see on the screen here, so this is approximately 10 to 15 credits. I did have around 170 credits when I did start this process. Just to give you an idea of all of the different things that it does from the initial generation, the planning through the testing, that's about 10 to 15 credits for what I had shown you here. Additionally, you can sync this to GitHub so your code isn't trapped in the environment. You can take this code and go ahead and deploy it wherever you'd like. And additionally, there are some other cool features a part of the platform like being able to fork this or being able to enable Ultrathink if you are a part of one of their pro paid tiers here. And last but not least, if you do want to deploy applications to the platform, one neat aspect of the platform is you will be able to host them directly on the platform for 50 credits per month just to circle back to the cost of it. So effectively, you're going to be able to host your application for half of the credits of what comes within the standard plans. And now in terms of actually hosting a production application, that is quite reasonable. So 10 bucks, you don't have to worry about provisioning all of the different services or worrying about maintaining that, that will be all on the team over at Emergent. Just being able to access all of these capabilities while being able to sync with GitHub definitely does make it a pretty compelling option in my opinion. But otherwise, that's pretty much it for this video. If you found this video useful, please like, comment, share, and subscribe. Otherwise, until the next