Website import
Turn a public website into structured sources — pages, images, and brand evidence.
A lot of organizational knowledge is already published — help centers, product pages, public policies. Website import turns it into sources without copy-pasting.
How it works
Give Vidext a public URL or domain. It then:
- Maps the site — discovers the site's pages, including subdomains belonging to the same company.
- Selects what matters — ranks the discovered pages and picks the most relevant ones, rather than ingesting everything blindly.
- Imports the selection — each chosen page becomes a source document: clean text plus the useful, non-decorative images.
Imported pages enter the same pipeline as uploads — processed into canonical text and indexed into the module's concept wiki. See The knowledge layer.
Two distinct jobs
Website content serves two purposes, and Vidext handles both:
- Knowledge — what the pages say becomes grounding for planning and generation.
- Brand evidence — what your site looks like feeds your branding profile, so generated modules match your visual identity.
Good candidates to import
- Your help center or knowledge base — the canonical "how it works" content.
- Product and pricing pages — for sales and support enablement.
- Published policies — for compliance modules grounded in the public source of truth.
Public pages only
Import reads what is publicly reachable. Content behind logins belongs in uploads or connected sources.
Last updated on