Talk:Kagi (search engine)

	This article was reviewed by member(s) of WikiProject Articles for creation. The project works to allow users to contribute quality articles and media files to the encyclopedia and track their progress as they are developed. To participate, please visit the project page for more information.Articles for creationWikipedia:WikiProject Articles for creationTemplate:WikiProject Articles for creationAfC
	This article was accepted from this draft on 7 February 2024 by reviewer Newystats (talk · contribs).

Websites: Computing

	This article is part of WikiProject Websites, an attempt to create and link together articles about the major websites on the web. To participate, you can edit the article attached to this page, or visit the project page.WebsitesWikipedia:WikiProject WebsitesTemplate:WikiProject WebsitesWebsites
???	This article has not yet received a rating on the importance scale.
	This article is supported by WikiProject Computing.

Merge information from draft

@StreetcarEnjoyer Please integrate information from Draft:Kagi (search) into this article. Greatder (talk) 18:56, 1 March 2024 (UTC)[reply]

Most of the information from the draft is here. I've added a bit of information about Orion Browser. Should more secondary sources arrive it has the potential to be spun off into its own article. StreetcarEnjoyer (talk) 21:13, 6 March 2024 (UTC)[reply]

@StreetcarEnjoyer I'll remove the draft then. Greatder (talk) 06:06, 7 March 2024 (UTC)[reply]

Poor citation for claims of not collecting user actions

Reference 8 (at the time of writing) supporting "The site does not collect user actions such as searches" is simply a link to a news website which quotes an employee of Kagi. It's effectively just PR-speak. As far as I'm aware Kagi has not proved or demonstrated that they do not collect such data, so why is Wikipedia stating it authoritatively on the page? It should either be removed or amended to say that the site claims not to collect user actions. 82.69.74.205 (talk) 17:27, 6 June 2024 (UTC)[reply]

Created Userbox

I've created a Kagi userbox for myself, (and anyone else who'd like to use it).

Code

Result

This user searches the internet with Kagi.

Usage

I added it to the list of website userboxes too.

Neuroxic (talk) 00:03, 11 November 2024 (UTC)[reply]

Criteria for reviews?

The article has a section "Search results quality" that cites one comparison of Kagi to other search engines that is generally unfavorable. From a casual investigation, this seems unrepresentative of reviews of the search engine. Most assessments are more favorable to much more favorable towards Kagi than the one citation. But what I've found have been opinions of users rather than rigorous apples-to-apples comparisons.

The comparison cited is not dated as far as I can see which seems like a major drawback as I presume that as a new service Kagi is undergoing significant development. So, I am somewhat inclined to remove it. At a minimum I would be inclined to qualify the comment and add other assessments. But, as I said, I have not found rigorous reviews. I have not investigated Wikipedia's standards or norms for this sort of information so I am leaving it as is for now but wanted to flag it in the hope someone else has time/experience to edit this section.

Jreiss17 (talk) 20:46, 3 January 2025 (UTC)[reply]

I've taken a look at the source and the methodology is all over the place. the user seems to be expecting either too much or not enough of the search engines tested, including expecting answers out of literal textbooks, and expecting results that answer a different question than the one asked. This isn't to say that the conclusion is wrong, but really I expect every search engine tested in that source to perform quite poorly.

The Marginalia results in particular appear to be significantly biased by the author's selection of search terms (ie it does great if what you're into is specifically open source software).

[1] seems to hold an actual head-to-head but only between kagi and google, and it doesn't appear to have been written by a kagi superfan either. the results seems to be that it's on par with google results, failing in some areas and suceeding in others (which makes sense for an engine populated by google results in the first place).

[2] is much less in depth. it tries more queries but only offers a final summary and doesn't explain what works and what doesn't for each query, just overall. the summary is, however, much easier to cite:

In general, Kagi seems to perform better whenever Google has too much of an incentive to perform worse. Popular searches and searches where Google stands to make the most money from advertising also tend to yield the lowest quality results. In contrast, Kagi seems to do exceptionally well on these searches. On less lucrative searches, both engines have similar results.

The author also notes a weaker image search.

It may be a good idea to note that while the default results are as they come, customisability may significantly improve quality, which is a point noted by several of the sources already cited in the article.

seeing as

This user searches the internet with Kagi.

, i'm not gonna go ahead and add these sources myself as that would probably fall under COI. If someone else can review the sources provided and edit the article appropriately that'd be great.
themoon@talk:~$ 14:00, 5 February 2025 (UTC)[reply]