I’ve created an automation that streamlines content creation from research to publishing and social media posting. I’ll show you exactly how to set up this powerful system using Airtable, Perplexity, GPT-4, DataForSEO, and FLUX-1 to generate well-researched, fully formatted articles and social media content.
This automation takes a simple topic and transforms it into a published article with multimedia elements and ready-to-go social media posts. By leveraging these tools, you can save time and ensure consistent, high-quality content across all platforms.
I go through the entire process in the video below.
Setting Up Airtable and Perplexity API
The automation begins with Airtable, where you add topics for your articles. Once a topic is added and marked as “ready for outline,” the system triggers the research process using the Perplexity API.
To set up Perplexity:
- Create an account on Perplexity
- Go to the API section and copy your API key
- Add Perplexity to your scenario in Make
- Select a model that ends in “online” for access to live web data
The system prompt for Perplexity is crucial. It instructs the AI to create a summary of facts, figures, and information for a blog post without introductory text or numbered headings.
Creating the Article Outline
After gathering information from Perplexity, the system uses GPT-4 to create an article outline. The outline writer takes the researched information and structures it into a specific format, including:
- A maximum of six sections
- Combining parts of the analysis if necessary
- Following the inverted pyramid structure
You can then review and edit the outline in Airtable before proceeding to the next stage.
Article Generation Process
Once the outline is approved, the system moves to the “ready for article generation” stage. This process involves:
- Creating a JSON outline of the article sections
- Writing each section using GPT-4
- Adding media (YouTube video or AI-generated image) to each section
- Combining all sections into a complete article
- Creating an introduction and key takeaways
The system uses a router to randomly decide whether to include a YouTube video or an AI-generated image for each section, adding variety to the content.
Integrating Media Content
For YouTube videos, the automation uses the DataForSEO API to search for relevant content. It then selects a video URL and adds it to the article section.
For images, the system uses the FLUX-1 image generation model. It creates a simplified prompt based on the section heading and article title, then generates an image that fits the content.
Publishing to WordPress
After generating the complete article with media, the system publishes it to WordPress. This process includes:
- Uploading the featured image
- Formatting the article content with proper HTML tags
- Adding YouTube videos and AI-generated images throughout the article
- Including sources at the end of the article
The system uses the Auto Upload Images plugin for WordPress to handle image uploads efficiently.
Creating Social Media Content
Once the article is published, the automation creates summaries and posts for various social media platforms:
These posts are tailored to each platform’s specific requirements and best practices.
Cost Considerations
The overall cost of running this automation is relatively low:
- FLUX API: A few cents per image
- Perplexity API: Less than 1 cent per call
- DataForSEO API: A fraction of a cent per call
- GPT-4: A few cents per article, depending on length
While the automation uses a significant number of operations in Make, the overall cost remains quite low, especially considering the time saved and the quality of content produced.
