Skip to main content
Native API About ten minutes, most of it field mapping

Publish to Webflow

Vupie creates items in the collection you point it at, filling the fields you map, then triggers a publish so the article goes live on the domain. Your design system does the rest.

Setup

How you connect it.

What you need

A Webflow site API token and a blog collection whose fields we can map.

  1. 1

    Create a site API token in your Webflow project settings.

  2. 2

    Paste it into the site's Integrations panel in Vupie and pick the target collection.

  3. 3

    Map the fields: title, slug, rich text body, summary, main image, and any reference field for category or author.

  4. 4

    Send a test item, check it in the Webflow editor, then publish.

On publish

What Vupie fills in for you.

Not just the body text. The fields that decide whether the page is indexable and linked are set on the way in, because an article nobody can find is not published, it is stored.

Title and slug
Rich text body
Summary and meta fields you map
Main image, uploaded to your assets
Reference fields for category or author
A site publish, so the item is live rather than staged

Before you pick this route

Worth knowing about Webflow.

Mapping is the whole setup

Webflow collections are bespoke, so the one-time field mapping is where the work is. After that it is unattended.

Required fields matter

A required collection field with nothing mapped to it will block the item. The connection test catches this before the first real article.

Publish is triggered, not assumed

The item is created and the site publish is fired, so the page exists at a real URL and can be crawled and index-inspected.

The same on every destination

The integration decides where. It does not decide what.

Whichever route you pick, the article behind it went through the same pipeline. Choosing Webflow changes the plumbing, nothing else.

The full method

Researched before written

Primary sources retrieved and tiered, your own numbers computed from your data, and the two or three questions only you can answer sent by SMS or Slack. The model may cite only from that list.

Reviewed by another vendor

Every draft is checked by a model from a different vendor than the one that wrote it. Sections below threshold are rewritten up to twice, and what still fails is refused rather than published.

Images, links and schema

Images with provenance metadata, internal links into your existing pages, and Article schema, all set through the integration rather than left for you to add.

Index inspection after publish

The URL is checked on days 2, 7 and 14 to confirm it was actually indexed, and AI-visibility tracking watches whether ChatGPT, Perplexity and Gemini start citing it.

Connect it once. Then it publishes without you.

Setup is about ten minutes, most of it field mapping, and it is the last time you have to touch it. After that, articles land on the calendar you approved.