It's Like the Early Days of Google!

Tuckerfan · Feb 19, 2024

Remember when the Google home page was just a simple search box, and you tended to get more of the results you wanted? This search engine does.

This Guy Has Built an Open Source Search Engine as an Alternative to Google in His Spare Time

Using Google has started to feel worse over the last few years, as results are seemingly taken over by SEO'd content, AI-generated results, and websites with tons of affiliate links and ads. As a response to this state of affairs, a single coder has launched a new, open-source search engine in part as a response to internet’s overwhelmingly corporatized and homogenous search ecosystem. The new search engine, called Stract, is running on a server in the basement of its developer’s office, is highly customizable and, based on feedback from users in the project’s Discord, is rapidly improving.

The project grew out of founder Mikkel Denker’s master’s thesis at the Technical University of Denmark, which was focused on helping people search their own files and documents, he told me in an online chat. He is set to finish that master's next week and will then pivot his attention to Stract fulltime.

“Most of our searches go through the same handful of entities (Google, Bing, Yandex),” Denker told me. “Even other search engines such as DuckDuckGo use Bing for their results. I found it very weird that there essentially is no way to browse the web in an open manner. So that's what I am trying to build.”

There are a handful of upstart independent search engines, and some of them are open source. But Stract feels particularly notable because it is highly customizable, runs fast, uses its own web index, and is highly usable despite being made primarily by one person in his free time over the course of the last few months (Denker says a friend provided a lot of initial help but has since gotten busy with a PhD program). Developer Rohan Kumar, who has been keeping an incredibly thorough and updated list of search engines with their own indexes for the last three years, wrote on their website that Stract is “my favorite generalist engine on this page.” Denker said that since being listed on Kumar’s site and getting some attention on Hacker News, the site has been handling about a thousand searches per hour.

Stract’s most interesting feature is called “Optics,” which are essentially customizable instructions that you can give to the search engine to help search for specific types of content or to narrow searches only to specific websites.

For example, there are currently Optics that allow users to only search websites that regularly appear at the top of Hacker News, to search only forums and message boards, to search only websites that are federated on the Fediverse, and to search only independent blogs. Because the project is open source, any web developer can write their own Optic or can take and modify any existing Optic (whose code is also open source). Stract also lets users “like,” “dislike,” or “block” specific websites, and lets them de-rank websites that have third-party trackers or lots of ads. The feature is similar to Brave’s Goggles, but Optics are more customizable: “The idea is to put as much control as possible over the search into the user’s hands,” Denker said.

“Part of Stract’s advantage comes from the fact that Google makes money from the webpages that contains ads from Google through their AdSense program, and are thus incentivised for these results to rank higher in the search results, whereas Stract can focus solely on which results are best for the user,” he said on the project’s Discord. Stract’s About page says “if we ever become evil (maybe by changing our motto) please take our code and start a competitor. The fact that you have this ability will make sure that our values will always be aligned with our users.”

Stract is currently not monetized in any way, but its website says it will eventually have contextual ads tied to specific search terms but that it will not track its users, which is similar to the system DuckDuckGo uses. Stract also plans on offering ad-free searches to paying subscribers.

As we’ve explored in several recent articles, Google has become increasingly frustrating to use over the last few years. I asked Denker whether he thought Stract could avoid being taken over by AI-generated and SEO-ed content. He said it’s a difficult problem to solve, but that his goal is to allow users to customize their own rankings with the use of Optics, “so everyone can decide for themselves what type of content should be prioritized.”

Stract is the latest entrant in a suddenly revitalized search market that includes Kagi, which is a paid subscription search engine I’ve been messing around with over the last month (and will be writing about the experience in more detail sometime in the next few weeks). I like Kagi so far, but the way it works under the hood (if such things matter to you) is as something of a search aggregator that combines results from Google, Yandex, Mojeek, Brave, and other large web presences like Wikipedia, Wikimedia Commons, TripAdvisor, and Flickr, to create its search results.

Stract, meanwhile, is creating its own index, which is a colossal undertaking. “The internet is a big place and most people know that it's not exactly trivial to build a search index,” Denker said. “With that said, it's actually also not as difficult as one would think.”

A search index is obviously critical to a search engine’s usability, and Stract felt a bit limited as I was messing around with it. I was unable to find a few specific websites I was searching for, presumably because they haven’t been indexed yet (though I was happy with the results for a handful of my searches). When I typed in “New York Times,” for example, the website for the New York Times didn’t show up. The good news is that Stract’s index is growing rapidly, and users have noted in the project’s Discord that their experience is regularly improving.

“I expect the search results will improve quite a lot in the next year. Currently Stract is still very much in beta, but as the index grows and as our ranking models improve the search results should follow,” Denker said. “The index is currently around 250 million pages, and as this gets bigger we will hopefully have better search results. I am especially trying to figure out how to best handle these navigational queries,” he added, referring to the New York Times search.

So far, Stract feels like something new and human in a space that has increasingly felt taken over by spammy results. "The best way to support Stract is to simply use the search and talk about it. Everything is entirely open source, so if people feel like it, it would also be a tremendous help to get some more contributors," he said. You may not be able to use Stract as your only search engine until its results improve, but using it helps the project and will hopefully make it better in the long run.

“The search on Stract currently runs on 2 bare metal servers, with one small 1u server bearing the majority of the load,” he said, adding that, for now, “It’s actually surprisingly feasible. And hey, if no one tries to do it, it for sure won't be done.”

“I rent an office desk here in Copenhagen where I can work, and they have a basement with some rack space I can use. I also rent some servers at Hetzner in Frankfurt, but right now most of it actually runs here in the basement,” he added. “Not quite a garage, but somewhat close.”
Click to expand...

You can try it out here.

NAHTMMM · Feb 22, 2024

For example, there are currently Optics that allow users to only search websites that regularly appear at the top of Hacker News, to search only forums and message boards, to search only websites that are federated on the Fediverse, and to search only independent blogs.
Click to expand...

Yes. This is what I dream of.

Log in or Sign up

It's Like the Early Days of Google!

Tuckerfan BMF

NAHTMMM Perpetually sondering

Log in or Sign up

It's Like the Early Days of Google!

Tuckerfan BMF

NAHTMMM Perpetually sondering

Useful Searches