[Hackbraten]

Until you realize that Kagi only works well because it uses a (paid) third-party API which behind the scenes does a classic Google search, scrapes its results in real time, throws out the ads, and then returns the cleaned-up results.

If Google Search changes, then Kagi's search will be impacted directly.

[BrunoBernardino]

This isn't entirely true, because they use more than one search index.

[Hackbraten]

The other search indexes are largely negligible in comparison: [0]

> This is not a competitive market. It is a monopoly with a distant second place.

> The search index is irreplaceable infrastructure. Building a comparable one from scratch is like building a parallel national railroad. Microsoft spent roughly $100 billion over 20 years on Bing and still holds single-digit share. If Microsoft cannot close the gap, no startup can do it alone.

[0]: https://blog.kagi.com/waiting-dawn-search

[BrunoBernardino]

Thanks for searching the source. It sounds to me like they are _not_ using Google nor Bing directly:

> With Google and Bing, we failed - not for lack of trying.

> Bing: Their terms didn't work for us from the start. Microsoft's terms prohibited reordering results or merging them with other sources - restrictions incompatible with Kagi's approach. In February 2023, they announced price increases of up to 10x on some API tiers. Then in May 2025, they retired the Bing Search APIs entirely, effective August 2025, directing customers toward AI-focused alternatives like Azure AI Agents.

> Google: Google does not offer a public search API. The only available path is an ad-syndication bundle with no changes to result presentation - the model Startpage uses. Ad syndication is a non-starter for Kagi's ad-free subscription model.1

And BTW, I think this is possible (perhaps, as someone suggested below, with smaller and niche-focused indexes). At Uruky we integrate with several independent search providers and indexes, and only one of them (Serper) uses Google's results indirectly — not the most popular (though it was added by pressuring demand). From our FAQ [1]:

> For web search, Uruky currently integrates Mojeek, Marginalia, EUSP (Ecosia/Qwant) (only works with French, German, or English), Linkup, Serper, and Uruky Site Search.

[1] : https://uruky.com/faq

EDIT: Sorry, I re-read carefully the whole thread and it seems we're saying similar things. Your claim, from the start, is about the SERP, not direct search, which is supported by their blog post as well:

> Because direct licensing isn't available to us on compatible terms, we - like many others - use third-party API providers for SERP-style results (SERP meaning search engine results page). These providers serve major enterprises (according to their websites) including Nvidia, Adobe, Samsung, Stanford, DeepMind, Uber, and the United Nations.

[AlienRobot]

This is a problem of their own invention.

Nobody said you have to index the entire web.

The web would probably be a lot healthier if we had several small search engines that focused on niches rather than 5 failed search engines that tried to index everything that was ever written and then ended up paying Bing.