This idea is either extremely smart or an extremely stupid—no in-between. What if your LLM *is* your search engine? How would you look like inside it? Forget about Perplexity, DeepResearch. What if LLM is your entire Google? Pagination, links and everything - just like the old days: no chat... UI, classic Google vibe. If you're unsure what's that mean, watch the demo video below first. We call it LLM-as-SERP (Search Engine Results Page).show more

Try it yourself at Fun things aside, why are we doing this? While implementing DeepSearch/RAG or any search grounding systems, a core challenge is determining whether a question needs external information or can be answered from the LLM's existing knowledge. A lot of systems are implemented with prompt-based routing like: System: - For greetings, casual conversation, or general knowledge questions, answer directly without references. - For all other questions, provide a verified answer with external knowledge. Each reference must include exactQuote and url.

This prompt approach fails in both ways - triggering unnecessary searches, or missing critical information needs. So what if we simply ran search anyway? We could make one call to a real search API and another to an LLM-SERP system. This eliminates the upfront routing decision and moves it downstream where we have actual results to compare - recent data from real search, knowledge within the model's training cutoff, and potentially some incorrect information. The next reasoning step can then identify inconsistencies and weigh sources based on recency, reliability, and consensus across results, which we don't have to explicitly code in—this is already what LLMs excel at. One can also visit each URL in the search results (e.g., with Jina Reader) to further validate the sources. In practice, this on-site-grounding step is always necessary anyway; never rely solely on excerpts from search engines, regardless real or fake search engines they are.

Read more about this experimental idea from the blog post below: Plus we also provide a SERP API hosted by us that you can play with. The API mimics a full SERP endpoint where you can define the number of results, pagination, country, language etc. So feel free plug into your own DeepSearch/DeepResearch implementations as a "search engine" to see any improvement there.

So in short, by using LLM-as-SERP, we kinda transform the binary question of "is this within the model's knowledge or not?" into a more robust evidence-weighing process. Finally, it's open source: We're eager to hear your feedback on this interesting approach. So make sure to let us know!

Too many people waste time checking their search rankings daily when they should instead invest in building an SEO moat that creates more permanent visibility.

then I can also get insight of how the LLM internally ranks my queries, my ASO starting point 😅

Interesting.. would this a good proxy for how well a site is optimised for GenAI SEO?

🎯

Go after enterprise if you can do it right

this so cool

JINAAAAAAAAA you could’ve atleast tried good ranking.