Web Query Language

Web Query Language is a query tool built to help AI stitch together scattered information from across the web into a single answer. A human types one question into a search box and clicks through the resulting links one by one to find an answer. An AI system, by contrast, can read many documents at once and piece them together, so there's no real need to hand it links in order. This tool is designed around that difference, aiming to construct answers by combining multiple sources even for questions that no single page can fully answer.

Think of it like asking a librarian not just where one book is, but to find several books on a topic and summarize them for you. That second request takes far more effort from the librarian. Kagi says the same is true here: without tuning the index structure ahead of time for this kind of question, the cost of re-searching the entire web every single time becomes unmanageably high.

This product is being prepared as the next step beyond the web search index Kagi already runs, and as of this writing it has not yet officially launched.

The article mentions it in a line like "the company's next product in the works is Web Query Language." The name might suggest a programming language like Python or SQL, but it isn't something humans write code in — it's the name of a product meant to help AI construct answers by combining multiple web sources. As of the article, it hadn't launched yet, so no concrete usage details were available.