I built a thing as fun side project at Urlbox.

It's a fresh perspective on the web. A way to explore over one million homepage screenshots. You can zoom, pan and click to navigate.

It's called One Million Screenshots.

It's built using data collected since the start of this year. We'll soon share that data in new API. It's also been fun to experiment with a few things we're considering adding to Urlbox's core API.

Here's the link:
https://onemillionscreenshots.com

One Million Screenshots

Explore the web's biggest homepage. Discover similar sites. See changes over time. Get web data.

@Jot This is great. There used to be a screenshot site way back when that I used extensively for a project. This is even better.
@lordmatt Thank you! Glad to hear that!
@Jot Very nice. How did you select the million sites? I'm surprised that my humble blog is part of it.
@roytanck Thank you Roy! I used page rank from a Common Crawl Web Graph release: https://commoncrawl.org/web-graphs
Common Crawl - Web Graphs

Detailing Common Crawl's Web Graph releases, the technology behind them, and how to use them.

@Jot Very nice and fun to use!
I find myself wanting to find meaning in the spatial proximity between screenshots, is there any?
It would be fun to be able to group things by color, theme, keywords, etc. I do realize it'd be yet more work ^^

@mikamtb great to hear!

I have attempted to make some meaning. The top ranked sites are in the middle. I then pull sites towards the corners of each layer based crudely on how they look.

Have experimented with other approaches. In the next version I will try to do some clustering based on content.