The connector reads three kinds of content from your Webflow site, and each kind is translated through its own section of the Translate page:
CMS Collections — every entry of every collection (blog posts, case studies, products…). All text fields are scanned, minus non-text fields and any fields you exclude in Settings → Field exclusions. This is the most fully automated content type: CMS content is structured data, so the connector can read all of it reliably.
Static pages — landing pages, about pages, and any other Designer-built page. The connector scans the page's own text elements (headings, paragraphs, buttons, links' visible text) plus any component property overrides placed on that page.
Components — reusable building blocks (navbars, footers, cards, menus…). The connector translates the component's definition, once, and Webflow applies that translation everywhere the component is used across the site.
Not scanned today: HTML embeds / custom code blocks, text that is part of an image or SVG file, image alt text, link URLs, and page SEO metadata (title tag / meta description). If a piece of text lives in one of those, it won't arrive in wxrks — see the next question for how to recognize these cases.