We’re (a group of friends) building a search engine from scratch to compete with DuckDuckGo. It still needs a name and logo.
Here’s some pictures (results not cherrypicked): https://imgur.com/a/eVeQKWB
Unique traits:
- Written in pure Rust backend, HTML and CSS only on frontend - no JavaScript, PHP, SQL, etc…
- Has a custom database, schema, engine, indexer, parser, and spider
- Extensively themeable with CSS - theme submissions welcome
- Only two crates used - TOML and Rocket (plus Rust’s standard library)
- Homegrown index - not based on Google, Bing, Yandex, Baidu, or anything else
- Pages are statically generated - super fast load times
- If an onion link is available, an “Onion” button appears to the left of the clearnet URL
- Easy to audit - No: JavaScript, WASM, etc… requests can be audited with F12 network tab
- Works over Tor with strictest settings (official Tor hidden service address at the bottom of this post)
- Allows for modifiers: hacker -news +youtube removes all results containing hacker news and only includes results that contain the word “youtube”
- Optional tracker removal from results - on by default h No censorship - results are what they are (exception: underage material)
- No ads in results - if we do ever have ads, they’ll be purely text in the bottom right corner, away from results, no media
- Everything runs in memory, no user queries saved.
- Would make Richard Stallman smile :)
THIS IS A PRE-ALPHA PRODUCT, it will get much MUCH better over the coming months. The dataset in the temporary hidden service linked below does not do our algorithm justice, its there to prove our concept. Please don’t judge the technology until beta.
Onion URL (hosted on my laptop since so many people asked for the link): ht6wt7cs7nbzn53tpcnliig6zrqyfuimoght2pkuyafz5lognv4uvmqd.onion
To answer your questions in order:
You could let people host their own as a method of scaling. But that limits it to geeks like us.
Use kubernetes and let it scale and pay for hosting on cdns.
A subscription-based model might be the only viable one, since ads will inevitably lead to a conflict of interest and voluntary donations are mostly a no-go. The problem is that people are so used to the notion that everything is “free” that many are convinced that online services should always be free and balk at the idea of paying for anything.
Personally I pay for Kagi which has been decent enough
Whats “decent enough” mean? I’ve been curious and you’re the only person I’ve known who pays for it.
I pay for it, the results are quality and the fact that my brain doesnt have to sift through ad results and can just look at the real data is so nice. Additionally, they have a large number of “lenses” which can change the scope of your search. For example, they have a lens for searching lemmy as well as lenses for the “small web”, which filters out all the results from massive corporate websites and gives way more personal project sites and the like.
All in all I’m a fan.
I never thought id pay for Kagi and that paying for a search engine was ridiculous. Then I kept seeing loudly positive feedback from reputable people in my circle and tried the trial.
I pay for it and never have the “I only ever use !g on duckduckgo” problem.
Sorting by web pages with least ad trackers is a cheat code to find old style websites with people sharing knowledge for knowledge’s sake rather than profit.
I mean, a search engine is literally the last thing on the internet I’d pay a subscription for. In a world where literally everything else nickels-and-dimes us for subscription service, search engines, torrent trackers, game modders who paywall their mods, and other kitschy non-essentials are literally the first things to get shuffled off the monthly budget.
If we weren’t in such a deep recession that I pay as much a week for my gas as I do my groceries, with rent and ACTUAL bills eating the majority of what’s left, I’d feel a bit differently; but if wishes were horses, we’d all ride. I literally had to start growing my own green rather than buying it, the economy’s so shit.
A huge part of that is that most people don’t consider privacy concerns to be a cost. All they factor into their evaluation is whether it costs them actual money.