Introducing Open Lovable - Re-Imagine Websites in Seconds

--- type: transcript date: 2025-08-08 youtube_id: O7CQBH3FDvo --- # Transcript: Introducing Open Lovable - Re-Imagine Websites in Seconds In this video, I am super excited to open source Open Lovable, which is a platform that allows you to take existing websites, select the style as well as preferences in terms of how you want the site to look. You can plug in a URL. It's going to grab the contents of the website, and then you can reimagine the site however you'd like. You can use basically any model that you want. Some definitely perform much better than others. My preference has been Kimmy K2 ever since I started to build this. And what it will do is, as the name implies, is it's going to build an application based on the context of your website. As soon as we send in the prompt, what it's going to do is we're going to grab the contents of the page leveraging the fire crawl API. While we're grabbing the contents in parallel, what we're doing is we're spinning up a sandbox on something called E2B. E2B is a really great platform where what it allows you to do is essentially create these environments where you're going to be able to read and write to a file system without having to worry about setting up something like an EC2 and worrying about scaling and all of that. They manage all of that type of complexity. And what you'll see is everything will start to stream in and the way that it's structured right now is it is set up to generate V applications. It will go through all of the different files and generation and just in a number of seconds we have a reimagination of the firecross site in this neo brutalist style. What I can do from here is just to show you some simple demonstration. I can say I want to change the background of the hero to be yellow. I'll go ahead and I'll send that in. And what it will do is it will determine the piece of context that it has to edit of all of those different files within here. We see that it did select the hero image. The goal with this was to really be able to have targeted edits for what you're asking for. And the other thing with this is the way that it's set up is say if you want to generate the initial generation of the site with Kimmy K2, you can do that. And then if you want to try out some really creative things like with 3JS or maybe a charting library or something that's a little bit more involved, you can switch over to GPT5 or Sonnet or whatever models you end up plugging into this. Now, just a couple aspects of the platform. What you're going to be able to do within this is download the projects. What we can do within here is you can go and you can export this. You can install everything and you can work on it within something like cursor or cloud code or whatever environment you prefer once you have the initial generation or you can go and continue on editing within this. Another thing that I just started to add support for was the ability to add in packages. Let's say I want to add in a charting library to show a pie chart on the right hand side of the hero. And then as soon as we send in that edit, what we'll see is it will create a new component for us. In this case, we have the pie chart.jsx file. We have our pie chart here that it generated with GPT5. We have the continuity of the overall look and feel of what the hero area looked like. And you can switch between different models or add in other models. Say if Gemini 3 comes out next week, you want to plug this in, we'll add in support for all of that. Also, if you want to experiment with other different models, you can definitely do all of that. Now, in terms of getting started, the repo is now live. All that you need to get started is you can clone down the repo, install everything. You just need to grab a few different API keys, a key from E2B, one from Firecrawl, and then you can optionally select the different LLM providers that you want to use. Personally, I have been enjoying leveraging Kimmy K2 on Grock, but if you do want to try out GPD5, you can do that. Additionally, you can plug in the models from Anthropic if you're interested. And then once you have that set up, you can go ahead and npm rundev. And now you have the environment locally. If you like these types of projects, please star the repo, like the video, share this, feel free to use it however you like. A shout out to the team over at E2B Grock, as well as basically everyone else that's built all the phenomenal technology that's really allowed for this. Most importantly, I just wanted to give a quick shout out to the team at Lovable for building an incredible piece of technology. I've always been interested in exploring how some of these platforms have built some of the technology that so many people have loved and appreciated. When I started this channel, he actually had something called GPT Engineer, which was actually one of the very first videos on the channel. To be able to see how the team over at Lovable have been able to accomplish what they have, it's been absolutely incredible. This was definitely a really fun one to work on. If you found this video useful, please comment, share, and subscribe. And I definitely love a star on the repo if you enjoy these types of projects. I'll catch you in the next