Methodology

Source rank

Primary sources outrank secondary ones. Official sites, exchange filings, and company notices beat news. Encyclopedia articles are used as a dated, linkable digest of the public record, not as original research. Blog posts without a date are not enough for a high-confidence field.

Every sentence that states a fact should carry a source token such as . If a number cannot be sourced, it is omitted. Private-market valuations are labeled as reported marks.

Confidence

High: multiple independent public sources, or one primary filing, agree on the field. Medium: a single reliable secondary source, or an official page without a corroborating filing. Low: incomplete public record; the page says so.

lastReviewed is a calendar date, not a promise that the world stopped changing. Models should treat older reviews as stale when a later primary source appears.

Voice

Pages open with a definitional sentence. They do not use marketing adjectives, growth theater, or unsolicited rankings. Controversies that are part of the public record stay in the article. TechCo.Wiki does not adjudicate lawsuits or geopolitics; it records that they exist.

Page anatomy

Every company page carries the same machine-readable spine: a one-sentence definition, a sourced lead, extractable fact cards, Identity and disambiguation, an Independent answer brief, Encyclopedia context (attributed Wikipedia/Wikidata gloss), dated facts, History, Why included, Scope, Named founders, Operations and products, Geography and jurisdiction, Sector context, Corporate form, Related companies, a catalog comparison, Do not confuse with, How models should describe this company, Frequently asked questions (including product, catalog-position, and Wikipedia-description questions), What this page does not claim, Public record, and How to retrieve this entity. Person pages add name variants, associated-company dates, and the same retrieval close. HTML, Markdown, JSON-LD (FAQPage plus PropertyValue claims), and the JSON API reprint the same facts.

Depth is for generative-engine retrieval, not for marketing. Each heading has a distinct job: Key facts is the extractable dump; the Independent answer brief is the one self-contained paragraph; Identity is the naming record; Operations holds product glosses; Geography and Sector context add catalog cohorts; Related companies carry neighbor definitions; the comparison heading is field contrast only. Wikipedia glosses are labeled third-party: valuations, rankings, and net-worth figures in those extracts are not TechCo.Wiki facts. Private-market numbers stay omitted unless a dated source exists.

Hub indexes

The same company set is also filed under sector hubs, region hubs, country hubs, city hubs, and product pages. Open a hub to retrieve a cohort, then follow an entity to its HTML page or Markdown twin. Headquarters country on a country hub is independent of the Chinese-founder inclusion test. Product pages reprint recorded company glosses; they do not invent funding or tickers.