The Xavior
Discovery Model

A recursive web-discovery engine. Pages reveal people. People generate better searches.

Xavior always begins with web pages.

The web page is the initial observation — a site, a directory, a staff list, a LinkedIn profile, a licensing record.

Step 1
Find relevant web pages.
  • Chess clubs in Colorado
  • Lawyers in Chicago
  • Bridge teachers in California
Every page is stored as a source object.

URL · title · page text · discovery date · the search that found it · source type · extracted facts · links out · confidence signals.

Step 2
Extract businesses and people.

Owners, employees, teachers, attorneys, directors — with emails, phones, locations and social accounts. The page remains the evidence for every fact.

Step 3
Create or match entity nodes.

Email, social handle, phone, name + employer, name + city, domain. Xavior creates a new node — or attaches the observation to an existing one.

Step 4
Search again using the entity.

Sarah Chen · Fort Wayne Chess Club · Indiana · [email protected] — each field becomes a sharper new query.

The goal isn’t to fill fields.
It’s to find more pages.
Step 5
Repeat the loop.
SearchPagesEntitiesMore pages
The web page is the primary evidence layer.

A profile is a structured synthesis of every relevant page Xavior has found — with the exact source retained for each fact.

Xavior isn’t searching for fields.
It’s searching for pages that create context.
Pages reveal people.
People generate better searches.
Better searches reveal more pages.
Made with xavior.ai