Product guide

Websites

Public sites, blogs and CMS platforms — WordPress, Ghost, Webflow, Contentful and plain RSS — crawled and kept current.

This category covers anything published on the web — your own site, a CMS behind it, or feeds you want to watch. Six source types live here: Website / Help centre, WordPress, Ghost, Webflow, Contentful and RSS / Atom feeds.

All of them are read-only — none can write back to the site they connect to. The credential differs by connector: Website and RSS / Atom feeds need none at all, WordPress's is optional (blank still syncs published content), and Ghost, Webflow and Contentful each need an API key or token created in that platform's own settings. Every sync re-checks what's already there and brings in only what changed.

Choosing between them: if your site runs on one of the CMS platforms listed here, connect it directly — you get structured content, real authors and dates, and drafts where credentials allow, none of which a page crawl can see. Use Website / Help centre for anything else, including a site on a platform not listed here, or alongside a CMS connector to pick up a page the CMS itself doesn't produce. RSS / Atom feeds is for watching content outside your own site — competitors, industry news, regulator announcements.

The connectors

• OverviewIngest public web pages — a handbook site, a help centre, published policies — and keep them in step as the site changes.
• WordPressConnect WordPress to read the CMS itself — posts and pages as clean structured content with real dates and authors, drafts included when credentials allow. Deeper than crawling the rendered site.…
• GhostConnect your Ghost publication to make every published post and page searchable as clean text with its real publish date, authors and tags — deeper than crawling the rendered site, so “what did we…
• WebflowConnect Webflow to read the CMS itself rather than the rendered site — every item of every collection (blog posts, help articles, case studies, team pages) as clean structured content, drafts…
• ContentfulConnect a Contentful space to make every published entry searchable — the marketing site, help centre and product copy behind your site, with text and rich-text fields flattened to prose and each…
• RSS / Atom feedsAdd news and blog feeds — industry news, competitor blogs, regulator announcements, your own release notes — and every entry becomes a dated document tied to its feed's name, so the assistant also…

Connecting more than one source for the same site

A CMS connector and the Website connector can both point at the same domain without conflict — they're separate sources with separate documents. A common reason to run both: the CMS connector for structured content and drafts, and Website for a page the CMS doesn't produce, such as a hand-built landing page or a linked PDF.

Last updated 20 Sep 2026