Media database
Outlets, journalists and addresses, with their provenance attached.
The corpus is the product. Tens of thousands of European outlets and their newsroom contacts, deduplicated, enriched offline, and labelled with how confident the record is.
01
How do I narrow it?
- Country and language — where an announcement can actually land.
- Medium — daily and weekly newspapers, agencies, portals, magazines, television, radio.
- Subject — a controlled vocabulary at outlet level, so a fintech story reaches fintech titles rather than everything with a business section.
- Prominence — reach as a percentile of the corpus, with a country-relative variant that does not bury every national title outside the anglosphere.
- Address state — verified, unverified, pattern-derived or suppressed.
02
How does it avoid duplicating people?
A contact has one of two identities. With an address, the normalised address is the key. Without one — an impressum name, an RSS byline, a Wikidata person — the key is the name plus the outlet, and a name alone is never enough: an email-less record needs a full name, an outlet link, and corroboration such as a role title or a profile.
Rows that have an address carry the name-based key too, so an address discovered later merges into the same person rather than forking them into two.
03
How do I know an address is real?
Because the product tells you which kind it is. An address observed on a published page is not the same as one derived from an outlet’s pattern, and the second kind is labelled so it can be excluded in a single filter. Verification against a provider is an explicit, paid step; nothing is silently marked valid.
Scholarly journals are tagged rather than deleted — a legitimate target for a research announcement, and a mistake in a press audience — so an audience can drop them in one move.
04
What of this is public?
Outlet identity only, at the media directory: name, type, country, languages, founding year, owner. Journalist names, roles and addresses are never published, never in structured data and never in a sitemap. See the page for journalists.