Repurpose YouTube Videos into Social Media Posts Using RSS, Dumpling AI, GPT-4o, and Airtable

---

 Introduction

Tired of watching YouTube videos and manually turning them into social media content? This automation makes your job easier. With one click, you can:

• Detect new YouTube uploads automatically
  
• Extract video transcripts with Dumpling AI
  
• Use GPT-4o to turn transcripts into social posts for Instagram, Facebook, and LinkedIn
  
• Auto-generate platform-specific images using Dumpling AI
  
• Save everything into Airtable, including captions and images
  

This is the perfect solution for content creators, marketers, and agencies who want to repurpose long-form content into ready-to-post social media content across multiple channels — automatically.

Let’s walk through the process step by step.

---

✅ Step 1: Detect New YouTube Uploads via RSS Feed

Purpose:
Automatically trigger the workflow when a new YouTube video is uploaded.

Node: Get YouTube Uploads

What to configure:

• Feed URL: Use the RSS feed from your YouTube channel or RSS tool like rss.app
  
• Set up polling intervals to control how often it checks for new videos

Output:

Node:

Pro Tip:
You can use RSS feeds for playlists too, not just channels. This is useful if you want to curate content from specific topics.

---

✅ Step 2: Extract the Transcript Using Dumpling AI

Purpose:
Pull the full transcript from the YouTube video for processing.

Node: Get YouTube Transcript

How it works:

• Makes a POST request to Dumpling AI
  
• Sends the video URL from the RSS feed
  
• Returns a clean, time stamped transcript
  

Request Body Example:

{

  “videoUrl”: “{{ $json.link }}”,

  “includeTimestamps”: “true”,

  “preferredLanguage”: “en”

}

Output:

Node:

Pro Tip:

You can include multiple language options if your audience is global. Dumpling AI supports multilingual transcripts.

Step 3: Generate Social Media Captions and Image Prompts with GPT-4o

Purpose:
Transform raw transcripts into platform-optimized social media content.

Node: Generate Social Posts + Image Prompts

What happens:

• GPT-4o receives the transcript and acts as a content marketer
  
• It creates 3 types of posts:
  
  • Instagram: short and friendly
    
  • Facebook: conversational
    
  • LinkedIn: professional and value-driven
    
• It also generates a tailored image prompt for each platform
  

Prompt highlights:

• Asks GPT to identify core messages
  
• Output is returned in structured JSON for easy downstream use

Output:

Node:

Pro Tip:
Adjust the prompt to include tone, hashtags, CTA styles, or match your brand voice.

Step 4: Structure Captions and Image Prompts for Each Platform

Purpose:
Split out the GPT-4o output into clean, separate variables for downstream use.

Node: Structure Posts & Prompts by Platform

What happens:

• Separates each platform’s:
  
  • Caption (post)
    
  • Image prompt
    

This makes it easier to send them into their respective content and image creation flows.

Output:

Node:

Pro Tip:
Make sure your Airtable schema matches these fields for clean data mapping later.

Step 5: Generate Images for Instagram, Facebook, and LinkedIn (via Dumpling AI)

Purpose:
Create stunning AI-generated visuals tailored to each platform’s style.

Nodes:

• Generate Instagram Image
  
• Generate Facebook Image
  
• Generate LinkedIn Image (Dumpling AI)
  

How it works:

• Uses the generate-ai-image Dumpling AI endpoint
  
• Sends platform-specific image prompts
  
• Returns a public image URL ready to be attached

Node:

Pro Tip:
Adjust aspect ratios (1:1, 16:9, etc.) to match platform preferences. Instagram likes 1:1, LinkedIn can handle 4:5 or 16:9.

Step 6: Save Captions to Airtable for Each Platform

Purpose:
Store content in a structured, reviewable format.

Nodes:

• Save Instagram Caption to Airtable
  
• Save Facebook Caption to Airtable
  
• Save LinkedIn Caption to Airtable
  

What happens:

• Sends caption text and platform name to Airtable
  
• Creates a new row for each platform

Node:

Pro Tip:
Use views in Airtable to sort by platform, or add fields like “Approved” or “Scheduled” for content workflow management.

---

✅ Step 7: Upload AI-Generated Images to Airtable Rows

Purpose:
Attach the image created by Dumpling AI to the correct Airtable record.

Nodes:

• Upload Instagram Image to Airtable
  
• Upload Facebook Image to Airtable
  
• Upload LinkedIn Image to Airtable
  

How it works:

• Uses a PATCH request to add the image URL as an attachment field
  
• Airtable then stores and displays the image inside each record

Pro Tip:
Ensure the correct record ID is passed along. Use node connections carefully to avoid mismatching images and captions.

Conclusion: This Is Your New Social Media Assistant

You’ve just built a full pipeline that can:

• Spot new videos in real time
  
• Turn long-form content into short-form gold
  
• Customize captions per platform (not one-size-fits-all!)
  
• Generate visuals that match each message
  
• Log everything in a system (Airtable) that’s clean, shareable, and scalable
  

This is how you:

• Turn a 10-minute YouTube video into 3+ pieces of platform-ready content
  
• Save hours each week while improving consistency
  
• Automate the repurposing process end-to-end, not just the writing
  

---

Ready to go further?

Here’s how you can take this workflow to the next level:

• Add a filter to skip YouTube Shorts or low-quality videos
  
• Send new posts directly to Buffer, Metricool, or Publer
  
• Add Slack or Telegram notifications when new content is ready
  
• Track engagement back in Airtable by syncing post performance