Screenshots are hard. We made them boring.

Our story

The problem

In July and August 2025, I was developing an internal tool, at the core of which was taking screenshots of websites, extracting and processing data from those screenshots, and then using that processed data down the line in the tool.

The task of taking a screenshot looked simple: use one of the many browser automation libraries like Puppeteer or Playwright, spin up a headless browser, call a "capture" method, and there you have it.

Nothing could have been further from the truth.

Cookie banners would pop up here and there, preventing us from capturing the data we needed.

Ads on the pages would significantly skew the pages' overall look, and as a result, the data we extracted from them.

Some sites wouldn't load because of CAPTCHA.

Some sites would show different content based on the user IP, and I would constantly have to use a VPN to capture content intended for different GEOs.

Full page screenshots would miss many parts of the main content because those parts are lazy-loaded, and you need to scroll the page down for them to load properly.

Etc, etc, etc.

And there was no reliable, built-in way to combat most of these.

Anyway, after 6 weeks of full-time work, this internal tool was finished. And of those 6 weeks, I spent 4 weeks on the screenshot functionality alone. Again, what was meant to be as simple as calling a single "capture" method and taking like 15 minutes took me a full 4 weeks…

And even then, I would still sometimes encounter edge cases here and there that warranted fixing.

However, because in early September 2025 I decided to scrap the internal tool altogether for reasons not connected to taking screenshots, those edge cases were never fixed, and the entire screenshot functionality was put on a shelf…

…to be taken off the shelf only a month later.

The idea

After scrapping work on the internal tool and the entire project it was part of, I started looking for new projects to work on.

Previously, looking at what actual people are typing into search has proven effective for me when looking for new ideas, so I opened Ahrefs and started digging.

I identified several seed terms that, in my opinion, should have uncovered some new business ideas to work on: checker, api, and a few others.

"checker": This one yielded ideas like "index checker", "site availability checker", "ping checker", "ada compliance checker", "spam checker", and many other good ideas. The problem was that many of these had very low search volume, or some very strong, long-established competitors. And usually both.

"api": This uncovered multiple good ideas that I really considered going after:

"logo api": good enough monthly search volume, relatively easy to implement, manageable competition, clear marketing channels. But the legal side of providing other brands' logos as a service seemed controversial, at least according to ChatGPT, so I eventually decided to abandon this idea.

"odds api": very large volume and lots of related keywords, the competition is medium, but just like in the case of the "logo api" keyword, running this service would involve constantly scraping other websites, which I preferred to avoid.

"domain availability api": low volume and lots of "free" keywords, which would indicate that the demand for a paid service is not that high. However, I still considered going after this business idea as it was very easy to implement, and I used a similar service on another project of mine, so I had a good understanding of who the target customer is and how the service should work.

All the ideas above sounded valid, but none of them ticked all the boxes I was looking for.

And then I saw another keyword…

"screenshot api": decent search volume on the main term, some additional keywords with medium search volume as well. There are a lot of competitors, but the competition in search specifically is medium. And from my own experience with taking screenshots, I knew immediately that the pain point was real.

And I already had all the necessary parts of the "screenshot api" in place, developed only a month ago as part of another project that was later scrapped.

I decided not to look for better ideas, and started working on the screenshot api project.

On September 27, 2025, I registered the domain screenshotscout.com.

And on February 1, 2026, Screenshot Scout went live.

The story begins.