How to Use Make and ChatGPT to Automatically Populate Your Webflow CMS

Automating pSEO content creation with Make and the OpenAI API lets you generate and publish dozens of Webflow pages without any manual work. It’s the ideal pipeline for large-scale programmatic SEO. Sandro, cofounder of Gemeos Agency, walks you through how to build it.

Prerequisites

• A Make account (formerly Integromat) with a paid plan
• An OpenAI API key with access to GPT-4
• A Webflow site with a CMS collection and access to the Webflow API
• A Google Sheets file with your input data (pSEO variables)

1. Prepare your input Google Sheet

Create a Google Sheet with one row per page to generate. Each row contains the variables that will be injected into your prompt: city, industry, service type, target keywords. Add a "Status" column, left blank by default, that Make will update once the page is created to avoid duplicates.

2. Create the Make scenario

In Make, create a new scenario. Here’s the module structure, in order:

1. Google Sheets > Search Rows: fetches the rows where the Status column is empty
2. Iterator: processes each row individually
3. OpenAI > Create a Completion: generates the content with GPT-4
4. Webflow > Create an Item: creates the item in your CMS collection
5. Google Sheets > Update a Row: updates the status to "Published"

3. Configure the OpenAI module

In the OpenAI module, set up the prompt by injecting the variables from your Google Sheet. Example system and user prompts:

// System prompt
You are an expert SEO writer. You generate unique, programmatic content
optimized for search engines. Respond only in valid JSON, with no markdown.

// User prompt (with Make variables between double braces)
Generate content for a "Webflow Agency {{Ville}}" page.
Return a JSON object with these fields:
{
  "h1": "...",
  "meta_description": "...",
  "description": "...(120 unique words about {{Ville}})",
  "slug": "webflow-agency-{{slug_ville}}"
}

4. Configure the Webflow API module

In Make, add the "Webflow > Create an Item" module. Set up the connection with your Webflow API key (Project Settings > Integrations > API Access). Select your CMS collection. Map each Webflow field to the corresponding value extracted from the OpenAI JSON response.

To extract the values from the JSON returned by OpenAI, use the "JSON > Parse JSON" module between the OpenAI module and the Webflow module.

5. Handle errors and limits

Add an error handler on the OpenAI module: if the response isn’t valid JSON (which can happen sometimes), the scenario should log the error without crashing and move on to the next row. In Make, use the "Error Handler > Resume" function for this case.

Conclusion

A Make + ChatGPT + Webflow CMS pipeline takes 30 minutes to an hour to set up and saves you dozens of hours on your pSEO projects. Always keep items in Draft mode and check a sample before publishing at scale.

• Use case 1: an agency creates 100 local pages for a client overnight
• Use case 2: a SaaS generates integration pages automatically for every new partner
• Use case 3: a comparison site publishes new product pages as soon as a row is added to the Google Sheet