Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Ask HN: What search engines are you using?
65 points by t0bia_s on Nov 19, 2022 | hide | past | favorite | 58 comments
Can you share your recent experience with different search engines?

I've tried brave, duduckgo, kagi, start page, qwant, you, searX. Relevance of results is average. I want more organic results, to text made for content (blogs), not text made for clickbait (publishing sites). Add "reddit" to search words to get more organic results is also wired phenomenon.

I just feel that those search engines are more for advertising than searching relevant results. How you deal with that?

For example teclis.com is good for science and academic articles, are there any other search engines more relevant in different subjects?



I have been using Kagi for 6 months now, and really enjoy its clean interface and relevant results. Being able to boost the ranking of relevant domains and downgrade the low-quality ones is a game-changer!

To remove pinterest and slackoverflow mirrors from all search engines, I use https://letsblock.it/filters/search-results


Same. Due to the steady decline in the quality of Google results (even with an ad-blocker), I made the switch. Never thought I would pay for a search engine but my experience with Kagi has been absolutely worth it. I enjoy their product and am aligned with their philosophy so I’d love to see them succeed!


I moved from duckduckgo to kagi just because I like paying for a service directly when I can. Works great.

I am conscious of my true identity being linked to my searches when I use it, though.


Having to have an account and your credit card info, how can they claim to be private? Am I missing sth?


You're confusing private and anonymous. Anonymous would be that no one knows you searched "cucumber." Private would be that someone could know that you did but it isn't shared with every single person for any reason. DuckDuckGo for instance is private in the sense that no one knows you searched what you searched, but Bing Ads knows the query to serve you a relevant ad. Signed-in to Google, Google would know your query and target an ad based on it, while adding it to a profile— in other words, it isn't private or anonymous. In a perfect world or using a search engine like searx, it's private in the sense that only the searx server knows what you searched and anonymous in the sense that Google wouldn't know who from that searx instance searched a query.


Thank you, makes sense now. So anything that isn't anonymus can potentially get sold or obtained by the gov.


Brave search, which tries to pull in Discussions results, vs Google who originally had a fantastic discussion search feature, but it was killed to optimize the Ads business.

Brave has been better than DDG, and the only reason I use Google now is product & price search. Google’s shopping / product search is still unmatched for price comparison and retailer discovery, given many companies feed inventory data to Google for Ads.


I've been using Neeva [0] for a few months now. I'm paying for their Premium plan because I want to support alternatives to Google. A few times when I noticed bugs with Neeva I would contact them and surprise surprise a human would actually reply and we'd chat about the issue, and it'd get fixed in a week or so.

Before Neeva I'd been using DDG but with DDG I'd frequently find myself going back to Google for complex searches. So far with Neeva I haven't had to resort to Google almost ever.

I hadn't heard of Kagi until I saw this thread, but it seems Kagi and Neeva are competing in the same space. Neeva has a cheaper premium plan. I'm curious if anyone has tried both and what they see as the main differences.

[0]: https://neeva.com/


Google. I haven't spent much time with the others.

I use "reddit" a lot. I wish there was an option to only get results from private individuals on any site.

There's a big problem with sites that scrape content from other sites. They're named fiiwebbcjiriebdbjfjeb and the like, and aggregate crap. If you click them the actual page has nothing to do with the preview and is some kind of scam.

Sometimes they seem to have exactly what I was looking for but I can't find where it came from, and the site is just a preview plus a scam. I wonder if some are using GPT3 rather than scraping.

There's also a problem with less dishonest scraping sites that ate just annoying. Like those sites that make a summary landing page from a github repo and don't actually add much value. They just clog up results.

And of course, pintrest, which only has the thing you clicked on about half the time and yet seems very popular.


Google :)

It's fast and convenient, but there's no free lunch - one pays with personal data - Google always tries to pretend the concept of privacy doesn't exist.


DuckDuckGo but it's just Bing and the results suck.

Sometimes there's something I know exists and DDG can't find it, but Google does instantly. I'm considering switching back to G.


!g search produces google results


I did this for awhile with DDG. I had to do it so much I switched back to save myself typing those two extra characters 90+% of the time.


Makes sense. I must say I’ve seen an improvement in ddg results over time.

But, I have a separate browser instance for work where I’m logged in and use google for the extra benefit of customized results.


I like Kagi. I found that DDG is really bad and often resorted to !g; but I gave Kagi a shot, and it's become my main engine.


https://search.brave.com/

I've been loving it


Former journalist here. DDG Lite is my browser start page. In reality, though, "!g keyword" or even "!g keyword site:somesite.com" is still the most common query. Old habits die hard. Strangely, though, I have no problem using DDG on a console-only Linux system (with a text-only browser).

I complained about the degradation of Google's search results in a previous thread. Thankfully, a fellow HNer shared some really useful tricks to improve googling: https://news.ycombinator.com/item?id=32519252

In conclusion, adding "inurl:" and "intext:" parameters have really been a game changer -- to the better -- for me. I'm a fan of old-time, text-heavy Internet, for which phpBB-style forums are a goldmine. So "inurl:viewthread" or "inurl:viewtopic" is my most preferred way to make browsing fun again.

As just some guy from the countryside, I'm afraid of the sheer "mass" of Google as a company, all that surveillance capitalism, UI bloat, etc. Dislike the search results prioritizing they do. But, well, in practice, "inurl:" and "intext:" do work fairly well.

I also have a list of favorite knowledge-base style websites, databases, etc, for which I aim to write simple custom searching scripts, and combine the results in the end. Something like Julian Assange's Surfraw, only much simpler: https://en.wikipedia.org/wiki/Surfraw


>I also have a list of favorite knowledge-base style websites, databases, etc

Please consider sharing that list on an awesome-x style GitHub account if it is not confidential. It could be added to by others too.


I'm not using any search engines. Trying to browse the web using a combination of organically discovered links, portals, bookmarks and history. I'll want to write a plugin in the future that automatically caches the pages I visit, so I can have a local search engine of everything I ever visited.

Sometimes I use Wikipedia or Wiktionary to find generic terms.

That said... sometimes it doesn't work, and I have to cheat using DDG.


You mean something like scraping links in pages that you already visited? It's great idea. Visiting wikipedia pages could make list quite a long!


Currently testing out spot which was forked from searx:

https://spot.ecloud.global

https://gitlab.e.foundation/e/infra/spot


I've been trying https://you.com/. The UI can be a bit confusing, which is owed to the fact that they are actually trying something new.


I don't think I could trust anyone who can afford to buy a domain name like you.com. They're either so rich they've got money to burn on dotcom-era domain names, or they're misallocating what money they have... it doesn't sit well with me.


Same, search results are actually pretty good around 60-70% of the time. Specially around programming.


I’ve tried Ecosia, I want to use it, but the results are really hit or miss. Basic stuff, sure, but when you’re actually digging for info it falls short. I’m back at google but I wish I had found an alternative.


Face the same problem you're facing. Our small team at PsyTech believes that your problem isn;t really with Search, as much with Discovery. You should be able to fetch content on a pull basis without plugging in keywords into an input box everytime.

To solve all of this, we are trying to build mutter, as a better way to curate for everyone. Please do try it out and share your feedback@psytech.ai, and hopefully you won't have to Search again :)

https://mutter.cards

TS


Apologies, I didn't see a FAQ and I'm not going to create an account just to find out the answer; according to

> Streams are a collection of posts.

This is trying to be ... twitter? instagram? reddit? something other than a search engine that would enable one to "discover" the answer to arbitrary questions?


I made my own experimental search engine by scraping the edu domains on DMOZ, using an algorithm to summarize the most relevant paragraphs on each page, and then ordering the results by relevance. It is useful for finding experts and academic departments or programs: https://www.locserendipity.com/edu.html


Kagi almost exclusively.


Because they regard privacy as important am using Qwant https://www.qwant.com/ and Ecosia https://www.ecosia.org/


I use Google. It’s still the best. Overdoing it a bit on the ads, but they’re pretty easy to ignore.


https://ublockorigin.com/


you.com CEO here.

We are seeing a lot of growth and retention with developers on https://you.com/code You can vote on which search-apps you like and soon block or pin them completely. You will also be able to submit your own search-apps soon. We believe in an open platform for search.


https://seirdy.one/posts/2021/03/10/search-engines-with-own-...


Thanks, that is very useful list!


DDG except when I need to run a simple math "query" that somehow only Google allows to process (mostly when budgeting, 1+1+1+1+1= - DDG doesn't do anything, Google executes)


I'm cognizant that it's your workflow, and who am I to butt in, but surely there are local processes available to you that don't involve HTTP round-trips to a datacenter to do math? For example, Spotlight will do math, and then command-c will copy the result to the clipboard


That's very much my use case though, all my expenses are on my calendar and I add them all once in a while - so the source is a browser tab, I put the equation in the address bar as I'm reading it inline, then hit enter when I'm done and navigate away. Using a local app would actually add one more step to that process.


Not a search engine but pretty often I will ask GPT-3 a question now if it's not easily googled. Often gives the initial context to Google follow up questions.


I used duckduckgo for quite a long time until they proved to everyone they cant be trusted.

Went over to searx, but recently I was getting lots of empty results because they got banned or something.

Currently at Brave.


Another vote for Brave. On the rare occasion that Brave doesn't get the result I want, I'll go back to Google and bypass DDG completely now.

My experience is that Brave is better than Google probably 50% of the time, equal probably 40% of the time and then inferior 10% of the time, typically when I need location specific info where Google Maps + Reviews can't be matched.


I have similar experience. A lot of SearX searches have time out results - none. Neither for selfhosting SearX or .be. Brave is my currently as well but often cannot find what I'm looking for. I guess a lot of data I need are hidden in Discord chats.


Whoogle It's self hosted, and just forwards to Google. But, it's HTML only, with tons of options, and minimizes tracking. It even has an option to route through Tor!


If you self host, only use your instance yourself and don't use tor routing, how does it minimize tracking?

A single ip is still exposed, so they can just build profile out of that. That's why public searx instances are great, since its a single ip for multiple people.


Google. When I cannot find what I'm looking for, then Yandex (feels like using Google in 2006: little restrictions, no uncensored content)


I use Google, and I have a shortcut set up to append things to my search, like "-site:annoyingthinggthatalwayscomesupbeforetherealdocs.com"


LibreX

https://github.com/hnhx/librex/


I’ve been appreciating kagi mostly recently, that and DuckDuckGo are the only ones I currently use.


The apple springboard one - it is location aware enough. I use it the most. Otherwise it's ddg.


Qwant (Lite): I find it works best for French results, but it works pretty good for English as well.


it's just bing too :) I get ads from bing using that. (and maybe you should disclose something? :D


Indeed, disclaimer: I used to work for Qwant. But I've been using them even before ;)

It is not just bing though, Qwant has its own crawler and index and only fallsback on Bing when the results are not satisfactory.


Yandex feels like google classic.


DuckDuckGo, but I've gotten into the habit of !g almost all queries :(


Brave search: search.brave.com Occasionally ddg. Rarely google


Qwant, DDG, Brave and Mojeek.

EBG (Everything but Google)


shodan ;)

* https://www.shodan.io/


Searx.be




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: