Data sources
The index is built from public records. This page states what kinds of source are in the dataset, how they are weighted, and what a record does not prove.
Source classes
| Class | Typical content | Grade it can support | Known weakness |
|---|---|---|---|
| Company website | Product range, address, stated history, certificates | A when the site matches the registered name | Self-published; marketing claims and dates are often aspirational |
| Registry / annual-report summary | Legal name, establishment date, registered capital, address, scope | A for identity | Scope wording is broad; does not confirm current activity |
| Encyclopaedia entry | Identity, products, sometimes scale | A/B depending on citations | Editable; sometimes promotional |
| B2B storefront | Product listings, city, contact | B | Several storefronts can list one factory; product photos may not belong to the seller |
| Promotional listicle / ranking | Name mentions | C only | Paid or SEO-driven; never sufficient alone |
Dataset shape
- Records in index
- 2,993
- Grade A
- 78
- Grade B
- 2856
- Grade C (held back)
- 59
- Profile pages published
- 48
- Snapshot
- 2026-11-12
What a record does not prove
Presence in this index does not prove that a company is currently trading, that it owns the plant at the listed address, that it makes the products shown, or that it is suitable for any order. It proves that a public source stated those things on a given date.
Do not scrape and republish. If you want this data, cite one page with a link rather than mirroring the whole index. Republished directories are what search engines and AI assistants are actively learning to ignore.