MCP connectorsWebflow

Webflow MCP: connect Webflow to Claude and ChatGPT

Ask Claude or ChatGPT which blog posts are still drafts, what the About page says today or what visitors sent through the contact form this week. Then have the next case study saved as a CMS draft. It all comes straight from Webflow.

With Deployed OS everyone connects with their own Webflow login and picks which sites to share. The AI sees those sites and nothing else. It reads, and with Edit access it saves new CMS items as drafts. Nothing is ever published.

9 read actions · 1 write action · Works in Claude, ChatGPT and Copilot

In practice

What your team can do with Webflow in Claude and ChatGPT

Content

Blog drafts straight into the CMS

Write the post in the chat, confirm it, and find it as a draft item in the Blog Posts collection with the fields filled in. Someone publishes it in Webflow when it’s ready.
Marketing

An SEO check across every page

Every page with its SEO title and description, with the missing and overlong ones flagged. A to-do list before the next release.
Sales

Leads from the website form

What came in through the contact form this week, each answer by field. Follow up the same day.
Web lead

What’s live and what isn’t

Which CMS items are drafts, which have unpublished changes and which are live. One overview before you publish the site.
Examples

Ask it like this

  • “Which items in the Blog Posts collection are drafts or have unpublished changes?”

  • “What did people send through the Contact form this week?”

  • “List all pages with their SEO title and description, and flag any that are missing.”

  • “Read the About page and tell me what is out of date compared with the Team Members collection.”

  • “Which fields does the Case Studies collection have, and which are required?”

  • “Save a new draft in Case Studies for Acme Logistics with the text we wrote above.”

The connector

What the Webflow MCP connector does today

  • Readswebflow_list_sites

    List the Webflow sites you shared with the connection, with time zone, last publish date and custom domains.

  • Readswebflow_list_collections

    List a site’s CMS collections, such as blog posts, team members or products.

  • Readswebflow_get_collection

    Read a collection’s fields: label, type, whether each is required and the choices of an option field.

  • Readswebflow_list_items

    List a collection’s items, up to 100 at a time, with field values and status (published, draft, changed since publishing, queued or archived), or only what is live.

  • Readswebflow_get_item

    Read one CMS item with its name, slug, status, dates and field values, rich text as plain text.

  • Readswebflow_list_pages

    List a site’s pages with title, path, SEO title and description, and whether each is a draft, archived or a collection template.

  • Readswebflow_read_page

    Read the text of one static page in order, with images by their description.

  • Readswebflow_list_forms

    List a site’s forms with the page each one is on and its fields.

  • Readswebflow_list_form_submissions

    List what visitors sent through one form, up to 100 at a time, each answer by its field name. Submissions hold personal details.

  • Writeswebflow_create_item

    Save a new CMS item as a draft, only when you asked and confirmed the content. It is not on the live site until someone publishes it in Webflow.

Need it to do more? We extend connectors for the companies that use them.

Security

Who sees what

Reads, plus CMS drafts
The connector reads sites, CMS collections and items, pages and form submissions. With Edit access it saves a new CMS item as a draft. It never publishes, changes or deletes anything that already exists.
Only the sites people share
Each person connects their own Webflow login and picks which sites to share. The AI sees those sites and nothing else.
You choose who gets it
Admins decide which groups at the company can use the Webflow connector at all. Everyone else never sees it.
Every call logged
Who asked, which connector, when. Deployed OS shows reads and writes per connector and per person, and the log can be exported.
Setup

Set up the Webflow connector in three steps

01
A Workspace admin creates an app in Webflow under Settings → Apps & Integrations → App Development, turns on the Data Client and selects read access to sites, pages and forms, plus CMS read and write. Your company owns the app.
02
The admin adds the app’s Client ID and Client Secret to Deployed OS and chooses which groups get Webflow.
03
Each person connects with their own Webflow login and picks the sites to share. From then on they just ask.
Compared

Webflow’s own MCP server or Deployed OS?

Webflow runs its own MCP server. It is hosted by Webflow, you sign in with OAuth, and it reaches both the Data API and the Designer, so a developer or designer can work on CMS content and on the canvas from an AI tool.

Deployed OS is built for the whole company. One Webflow app that your company owns, each person’s own login, admins choosing which groups get what, and a log of every call. The same connection works in Claude, ChatGPT and Copilot, next to your other tools and your company’s skills and guardrails.

Their documentation: Webflow’s MCP server

Questions

Webflow MCP: common questions

Yes. Webflow runs its own hosted MCP server, built for developers and designers working on a site from an AI tool, the Designer canvas included. Deployed OS gives the whole company one Webflow connection next to your other tools, with each person’s own login, access per group and a log of every call.

No. It reads, and with Edit access it saves new CMS items as drafts once you have confirmed the content. Nothing is published, and nothing that exists is changed or deleted. If you want it to do more, we can extend it for you.

Yes. Deployed OS is your company’s own MCP, so the same Webflow connection works in Claude, ChatGPT and Copilot. People keep the assistant they prefer.

A Workspace admin, with a verified email address and two-factor authentication on their Webflow account. Webflow offers app development on some Workspace plans only, so check yours. People in your own Workspace can then connect.

The AI is told to quote only what you asked about, and every read is logged per person. Deployed is an EU company and fully GDPR compliant, and you choose where your data is stored, down to the city.

It depends on the size of your team. Get in touch and we will walk you through it.

More connectorsAll connectors
Let’s talk

See it on your own tools.

Tell us which systems your company runs on. We’ll show you what your team can ask, and what IT gets to control.

Poyan Karimi
You’ll hear from Poyan Karimi
Co-founder · Deployed AI