Exporting a Webflow site you don't own the Workspace for: what a 62-post CMS looks like from the outside

A published Webflow site is static HTML, and the CMS leaves marks in it. We ran an export against a public Webflow site (marketing.flowbase.co) on 31 October 2026 and read the files that came out: 62 posts, 21 pages, 577 images (~94 MB), and 229+ other CMS items that are not posts.

What the HTML tells you

• data-wf-site on <html> marks every Webflow page; data-wf-collection and data-wf-item-slug mark CMS items. The sitemap gives the list: /<collection>/<slug>.
• Nothing says which collection is the blog. Here two collections read like blogs. Our rule: names like blog/posts/articles/news, else the largest collection. It's a heuristic, so the scan lists the other collections instead of hiding them.
• Dates are a design decision: 33 posts had one, 29 help articles showed none, so only the sitemap's lastmod (last publish, not first) is left.
• The body is the largest .w-richtext element; -p-500 resized image copies map back to the original file.
• The same newsletter block was repeated on 16 pages, with its images each time. Short sections containing a form are now dropped.

What a visitor's view can't give you

Categories/authors/tags (CMS references), forms, e-commerce, memberships, code embeds, layout, drafts.

Full write-up with the checklist: https://decamp.app/notes/webflow-cms-export