How Sampson Built and Scaled Windframe to 16,000 Tailwind CSS Enthusiasts

Meet Sampson Ovuoba.

Sampson is a Software Engineer and Product Builder based in Lagos, Nigeria.

He has always enjoyed building developer tools, and is drawn to products that remove friction from developer workflows and make it easier to ship high-quality applications.

Right now Sampson is focused on Windframe. He is working full time on the project with a friend.

My current product came out of a frustration I kept running into while building user interfaces.

Tailwind CSS was still quite new at the time, and I enjoyed working with it. I loved its defaults and consistency, but I still felt there was an easier way to build these interfaces. Translating visual ideas into clean, repeatable code took too much time and mental effort.

I wanted a way to design visually without losing ownership of the code. Something that respected how developers already think and work.

That idea became Windframe.

Windframe is an AI visual builder for Tailwind CSS that generates clean, production-ready code. You design visually and keep full control of the output. It is aimed at developers and designers, whether they have little or extensive knowledge of Tailwind, who want to generate high-quality UIs quickly without getting too deep into the code.

Windframe is an extremely hard product to build.

I started working on it in January 2021, and the first usable version took over seven months to ship.

I rebuilt large parts of the app multiple times because they did not work as expected. Along the way, I had to learn much more about how browsers work underneath and build a custom rendering engine to handle different types of code and design. It took many failed attempts to arrive at something stable.

I’ve seen many people try to build tools like this and give up after a few weeks or months. It’s easy to create a quick demo, but once real users start using the app in unexpected ways, the complexity becomes overwhelming. That’s when you realize how deep the problem really is. Getting from “this looks cool” to “developers trust this in production” is where most people stop.

I launched on Product Hunt on July 2021 and shared the release on Reddit as well. Within about two days, Windframe had reached its first 100 users. Conversion was low at first, but the signal was strong. Developers were interested!

I got the first paid customer on launch day from Product Hunt and I still remember seeing the first $16 monthly subscription come through on launch day. After months of work, someone had decided the product was valuable enough to pay for. That moment felt really incredible.

In 2021 when no code was blowing up, I let some of that influence Windframe and built for the wrong audience. I later removed most of those features.

As mentioned I posted about Windframe on the developers subreddit on Reddit and it went kind of viral. I got a lot of helpful feedback and even new paying customers from that post.

That was when I finally understood the main audience for Windframe was developers. I used a lot of the feedback from that thread to improve Windframe, and that still shapes what Windframe is today.

Some things clearly compounded over time.

Product Hunt and Reddit at launch

Both Product Hunt and Reddit helped get Windframe in front of the right audience early and also get the first 100 users.

Direct user feedback
Emailing users (example below) early and asking for feedback exposed real pain points. Those conversations consistently revealed more than analytics ever could. I still do this today.

SEO and developer-focused content
Features on the Svelte blog, Dev.to, and similar platforms created steady, long-term traffic.

Improving onboarding
Early onboarding dropped users into a blank canvas. Adding guidance (example below) significantly improved signups and conversion.

Some efforts simply didn’t pay off.

Facebook groups
Most groups were spam-heavy and brought low-quality users who didn’t convert or stick around.

Paid ads
Facebook ads were too expensive. I was acquiring users at roughly 1.8x their lifetime value, which is unsustainable for a bootstrapped product.

Building features without validation
Several features I thought users would love saw little usage and were eventually removed.

Today, Windframe has over 16,000 users and receives around 9,000 visits per month. It’s used by developers across different teams and countries to build real production interfaces.

I’ve seen Windframe featured on several blogs and developer publications over the years. At some point, a firm even reached out with interest in investing, but I chose to keep Windframe fully bootstrapped.

I’ve had developers from companies like a16z (Andreessen Horowitz) use Windframe for internal tools, which was surreal to see.

There have been many smaller wins that meant a lot. Developers sharing Windframe in communities. People recommending it in threads when someone asks for Tailwind tools. Users sending emails about how much time it saves them. Returning customers who have been using it for years.

Windframe has grown slowly through word of mouth, community posts, SEO, and consistent iteration based on user feedback.

AI can now generate UI faster than ever, but speed isn’t enough. The real challenge is creating high-quality, consistent, and polished interfaces.

Windframe doesn’t just generate layouts, it helps create stunning, visually consistent UIs. It combines AI with the visual editor to make building great designs faster and easier, with built-in guardrails that guide quality.

Please note that Sampson has also started Brandframe (beta), a spin-off from Windframe. I It uses an agentic harness to extract a company's design system from their URL and generate new components that align with their brand guidelines. You input a website, it analyzes the styles, then creates on-brand designs as needed.