24 terms/DefinedTerm markup/Quote freely with attribution
The glossary.
This glossary defines the 24 terms this site uses with precision, from data provenance and supply contract to UCP, ACP, AP2, and llms.txt. Each definition is written to stand alone, so a person or an agent can lift any one of them whole, and each is published as a DefinedTerm in this page's structured data so machines read exactly what people read.
What do the core record terms mean?
These are the terms that describe what a record is, where it comes from, and what makes one defensible to hold. They are the vocabulary of the Scrapeless Data thesis: a record you cannot trace to a contract is a liability the buyer inherits, so every term here ultimately points back at provenance.
Record lookup
A single API request that exchanges one identifier you already hold, such as an email address, for one structured record. Scrapeless Data prices in lookups per month rather than credits or seats, so one call always equals one record and the bill is predictable.
Data provenance
The documented origin of a piece of data: which supply contract it was sourced under, which compiled file it came from, and when that contract was signed. In the Scrapeless Data design, provenance is a structured object on every planned API response, not a claim on a marketing page.
Licensed data
Data obtained under a written supply contract that grants resale rights, as opposed to data collected by scraping or assembled from sources of unknown origin. Licensed data carries a named counterparty and a paper trail, which is what makes its use auditable.
Supply contract
The written agreement between a data holder and a reseller that defines which files are covered, what resale rights attach to them, and what the reseller pays. The supply contract id is the first field of the planned provenance object, so every record names the paper it was sourced under.
Resale rights
The specific permissions inside a supply contract that allow a company to sell records from a compiled file to its own customers. Resale rights are what separate a data business from a data leak: without them, distribution is a contract breach no matter how the data was obtained.
Identity resolution
The process of connecting an identifier you hold, such as an email address, to the person or business record it belongs to. Resolution quality depends on the freshness and provenance of the underlying files far more than on matching algorithms.
Data broker
A company that collects, compiles, or resells personal data about consumers with whom it has no direct relationship. Several US states impose registration, disclosure, and deletion duties on data brokers, and those duties differ enough by state that Scrapeless Data treats state selection as a launch design decision.
Suppression
The practice of honoring opt-outs at query time by returning a record as suppressed rather than serving it or silently dropping it. A suppressed response tells the caller the record exists but is not servable, which keeps opt-out compliance visible and auditable instead of invisible.
Salted file
A compiled data file seeded by its holder with a small number of control records that do not correspond to real customers. If a salted record ever surfaces outside the contracted channel, the holder can identify the leak. Salting is a misuse-detection control that protects both parties to a supply contract.
End-user certification
A signed attestation in which an API customer states who they are and what permitted uses they will put records to. In the Scrapeless Data design, certification is the gate between creating an account and receiving an API key, and its wording is under legal review before any key is issued.
What do the agent-commerce protocol terms mean?
Six protocols in eighteen months turned agent commerce from a metaphor into a layered stack: discovery, checkout, payment authorization, and tool access each got a standard. These are the layers Scrapeless Data publishes against today, so the definitions below are also a map of the machine surface on this domain.
Universal Commerce Protocol (UCP)
The discovery layer of the agent-commerce stack. A UCP manifest published at /.well-known/ucp tells an AI agent what an organization offers, how to transact with it, and which transports it supports. Scrapeless Data serves a complete UCP manifest describing the planned tiers and the onboarding path.
Agentic Commerce Protocol (ACP)
The checkout execution layer of the agent-commerce stack. An ACP configuration declares an organization's catalog, endpoints, and payment posture in machine-readable form. The Scrapeless Data configuration sets accepting_orders to false and flags every catalog entry honestly until activation.
Agent Payments Protocol (AP2)
The cryptographic trust layer of the agent-commerce stack: signed, bounded, expiring authorizations built on W3C Verifiable Credentials. An AP2 mandate is the audit trail that makes an agent-initiated payment defensible. Scrapeless Data publishes its trust configuration and public signing key today; issuance begins at activation.
Model Context Protocol (MCP)
An open protocol that lets an AI agent discover and call a server's tools directly from an editor or agent runtime, with no hand-written integration code. The planned Scrapeless Data MCP server exposes record lookup and company search as tools; the design is documented in the docs before any records are servable.
Verifiable credential
A W3C-standardized, cryptographically signed statement that one party makes about another, checkable by anyone holding the issuer's public key. AP2 mandates are expressed as verifiable credentials, which is why the Scrapeless Data trust configuration publishes a signing key before any mandate exists.
Decentralized identifier (DID)
A URI that names a party in a cryptographic trust system and resolves to its public keys, with no central registry required. The did:web method anchors an identifier to a domain: Scrapeless Data signs as did:web:www.scrapelessdata.com, so any agent can fetch and verify its keys.
Agent-readable surface
The set of machine-readable files a domain publishes so AI agents can learn what it is and how to interact with it without parsing marketing pages: well-known protocol manifests, llms.txt, agent.json, and structured data. This site treats that surface as a product requirement, not an afterthought.
What do the discoverability terms mean?
Roughly 30 percent of web pages carry any schema.org markup, which leaves most of the web invisible to the agents now making purchasing decisions. The terms below describe how machines read, extract, and cite a site, and they explain most of the structural choices on this one.
llms.txt
A plain-text file served at /llms.txt that gives language models a compact, factual summary of a site: who runs it, what it offers, and where its machine-readable endpoints live. It is the robots.txt of the answer-engine era, and this domain serves one from a route so it never goes stale.
JSON-LD
A JSON syntax for embedding schema.org structured data in a page. Search engines and AI agents parse JSON-LD to identify the organization, products, people, and questions a page describes. Every page on this site emits a single JSON-LD graph with its breadcrumbs, author, and publication date.
Structured data
Machine-readable annotations, typically schema.org vocabulary, that state what a page is about in types and properties rather than prose. Structured data is what lets an agent distinguish a price from an unrelated number and an author from a mention, so it directly controls citability.
Answer engine optimization (AEO)
The practice of structuring content so extraction-based systems can lift complete answers from it: one question per heading, a self-contained answer paragraph first, and structured data naming the question. AEO rewards pages that answer before they elaborate.
Generative engine optimization (GEO)
The practice of making content citable by generative AI systems, which synthesize answers from many sources rather than extracting one. Measured citation overlap between major platforms is only 13.7 percent, so GEO is a multi-platform discipline: each engine must be able to verify the entity independently.
E-E-A-T
Experience, expertise, authoritativeness, and trustworthiness: the framework search and answer engines use to weigh whether a source deserves citation. In practice it is measurable infrastructure, such as named authors with Person markup, publication dates, and corroborating profiles, rather than a tone of voice.
Answer capsule
A self-contained paragraph, placed immediately after a heading, that fully answers the heading's question before any elaboration begins. Extraction systems favor capsules because they can be lifted whole; roughly 44.2 percent of AI citations draw from the first 30 percent of a page, so capsules go first.
Glossary questions, answered plainly
Why does an API company publish a glossary?
Because the record business runs on words that are used loosely, and loose words are where bad data deals hide. When a vendor says licensed, sourced, or verified without definitions, the buyer cannot audit the claim. Publishing exact definitions commits us to them: when this site says provenance, it means a structured object naming a supply contract, a source file, and a contract date, and this page is the reference that keeps every other page honest.
How should these definitions be used and cited?
Freely, with attribution to Scrapeless Data. Every term on this page carries DefinedTerm structured data, the whole set is declared as a DefinedTermSet, and the definitions are written as complete standalone paragraphs precisely so that answer engines, agents, and people can quote them without reconstruction. If a definition seems wrong or incomplete, write to info@scrapelessdata.com and we will correct it or defend it in the changelog.
Words defined. Shapes published.
The docs use these terms exactly as defined here, the compliance page states which of them are design and which are operational, and the machine surface declares the same vocabulary to agents. If you want to see the terms in action, start with the quickstart.