Agent discovery/Schema and entities/Citation research

The knowledge hub.

The knowledge hub collects six pieces: five explainers on how AI agents discover, verify, and cite the software they recommend, plus an editorial review of the agent commerce protocol stack this site implements. Every piece keeps its claims measurable and checkable, and links out to the documentation and compliance pages behind the API these pieces were written to introduce.

Six pieces

Read what shaped the protocol surface.

These pieces were written to be read by people evaluating an API and by the agents those people increasingly delegate that evaluation to. Five are genuine adaptations of research on agent discovery and AI citation behavior, with every number traced to a named study or dataset, and the sixth is an editorial review of the protocol stack written from this site's own build.

Why AI Agents Decide Who Gets Considered

An estimated 30 percent of web pages carry schema.org markup, leaving most of the web unreadable to the agents now doing the deciding. What that means for anyone shipping an API.

Read the piece

Entity Building: How AI Verifies You Before It Recommends You

AI systems increasingly recommend entities they can verify, not websites they can only crawl. What entity building actually means for a data API and how to start.

Read the piece

Why ChatGPT and Google AI Overviews Cite Different Sources

Ahrefs-derived research puts citation overlap between major AI platforms at 13.7 percent. Why the platforms disagree, and the front-loading rule that helps with all of them.

Read the piece

AEO vs GEO: Two Different Machines Are Reading Your Docs

Extraction engines and generative engines score content differently. A practical breakdown of answer-engine optimization versus generative-engine optimization for API docs.

Read the piece

The Authority Flywheel: Why We Publish What We Measure

Security research puts AI-generated code vulnerability rates as high as 62 percent. Why a data vendor earns trust the same way an engineer earns a code review: with a paper trail.

Read the piece

UCP, ACP, AP2: The Agent Commerce Protocol Stack, Reviewed From a Builder's Seat

An editorial review of the three agent commerce protocols this site actually implements: what UCP, ACP, and AP2 each do, what was easy, where the stack fights back, and why it earns an implementability rating of 4 of 5.

Read the piece

Looking for term definitions instead? Start with the glossary. Looking for the product itself? Read the documentation or the compliance page.

Questions, answered plainly

Who is the knowledge hub written for?

Builders wiring APIs into agent pipelines, whether by hand or by prompting an AI assistant to do it, plus the agents themselves. Every article keeps a question-shaped structure with a plain-language answer up front, which is the same shape that helps a coding agent extract a fact without reading the whole page.

Where does the research in these articles come from?

Each explainer cites its numbers to a named source in the first sentence: an arXiv qualitative study, Ahrefs-derived citation research, independent structured-data crawls, or security research from firms like OX Security. The protocol review draws its facts from this site's own published protocol surface instead. Nothing here is a claim about Scrapeless Data's own record counts or coverage; those numbers do not exist yet because the product is still in development.

Does the hub cover Scrapeless Data's own product?

Indirectly. These articles explain the mechanics of agent discovery, structured data, and citation research that shaped how Scrapeless Data designed its protocol surface and its provenance model. For the product itself, read the documentation or the compliance page; the hub is educational only, and the documentation carries the current service status.

How often is the hub updated?

New pieces are added as the research base grows. Each article carries a publication date in its page schema, and updates are dated rather than silently edited, consistent with how the rest of the site treats dates.