Wikexa is a structured knowledge API for AI agents. It exists because the cheapest way to answer a factual question should not be to download a web page.
An agent asked "who runs Apple?" typically fetches the encyclopedia page for it. That page is roughly 15,000 tokens of HTML: navigation, edit links, citation scaffolding, infobox markup, references. The answer inside it is about five words. The agent pays for all of it, on every call, and throws nearly all of it away.
The source APIs are better but still not built for this. They return raw markup — templates, transclusions and markup that a model has to interpret before it can read a fact.
We do the parsing once, ahead of time, and keep the result in the shape an agent actually wants: templates resolved, references stripped, infoboxes turned into key–value pairs, sections split into an array, and structured claims mapped from property codes to readable names. What an agent receives is the part it was going to extract anyway.
Everything is prepared in advance and served from close to the caller, so a request does not wait on parsing, on a query being planned, or on a round trip to a machine in another continent. Warm lookups come back in single-digit milliseconds — fast enough to sit inside an agent loop rather than beside it, and efficient enough that ordinary use can stay free.
A tailer follows the upstream change stream and rebuilds records within minutes of an edit. That is the one thing a language model cannot do for itself: know what changed after it was trained.
More sources, and deeper coverage of the ones already in. Reference works, public-domain books, government and legal corpora, and scholarly metadata are all on the list. If you need something specific, say so — the roadmap is mostly driven by what people ask for.
Wikexa is built by the same people as cabrini.ai, a market-data API for agents. Same architecture, different audience. Wikexa is independent and not affiliated with any source it indexes.