SiteCreator

Inspiration

We noticed that a lot of local businesses have outdated websites, or no website at all—even though they offer great services. Making a website from scratch can take a lot of time and technical knowledge. We wanted to build something that helps people find these businesses, quickly create a modern website for them, and publish it online.

What it does

SiteCreator is an AI-powered tool that helps users find local business leads and build websites for them.

A user can search for businesses by city or category, choose a business, generate a custom website, make changes by typing prompts to an AI assistant, preview the website live, and publish it to GitHub and Vercel.

How we built it

We built the frontend using Next.js, React, TypeScript, and CSS. For the backend, we used Express, TypeScript, MongoDB, and Mongoose.

The backend creates separate Next.js projects for each website. It also starts development servers on different ports so users can preview their websites while editing them. We used node-pty to manage terminal processes, GitHub CLI to publish projects to GitHub, Vercel CLI to deploy websites, and an AI editing tool to help make changes from natural-language prompts.

Challenges we ran into

One major challenge was getting all the different parts to work together. The app has to generate a site, start a server, find an available port, wait for the preview to load, apply AI edits, and deploy the project.

We also had to handle login and authentication for tools like GitHub, Vercel, and the AI editor. Making the process feel simple for the user while a lot was happening in the background was harder than we expected.

Accomplishments that we're proud of

We created one platform that goes from finding a business lead to launching a real website.
We made websites that users can actually edit, not just view as a mockup.
We added AI editing so users can change content, colors, and layouts with simple prompts.
We made it possible to publish a site to GitHub and deploy it on Vercel with very little effort.

What we learned

We learned a lot about full-stack development, especially how frontend and backend systems communicate. We also learned how important it is to manage processes, loading states, and errors when an app is doing many tasks at once.

We learned that AI is more useful when it is part of a clear workflow. Generating code is helpful, but users without experience also need a simple way to preview, edit, and publish the result.

What's next for SiteCreator

Next, we want to add more website templates for different types of businesses. We also want to improve lead searching, make the AI edits more reliable, and add more customization options.

In the future, we would like to add client collaboration tools, analytics, and an easier way for users to track the websites they create and launch.