Introduction – The Use Case
Many organizations today maintain active websites that publish new content regularly such as news updates, job posts, or product announcements. However, distributing this content manually across multiple social media channels takes significant time and coordination, especially when each post must be formatted differently for various platforms.
To address this challenge, I created an automated social media publishing workflow using n8n, OpenAI GPT-3.5 turbo model and the Meta Graph API.
The goal was simple. Whenever a new article is published on the organization’s website the workflow should automatically collect it, reformat it into engaging social media post, and post it to multiple Facebook pages and an Instagram accounts all without human involvement.
This POC demonstrates how easily the daily work of a social media manager can be automated while maintaining brand voice and consistency.
The Workflow Overview
The workflow continuously monitors the website RSS feed for new articles. Once a new article detected, it retrieves the it, extracts essential fields like title, description, and image, and generates a fresh caption with hashtags using GPT-3.5 turbo. Finally the system posts the content to multiple Facebook pages and Instagram accounts using the Meta Graph API.
Node by Node Walkthrough
This Node by Node Walkthrough explains how the workflow runs from start to finish. It outlines what each node does and how the data moves through every stage
1. RSS Feed Trigger
The RSS Feed Trigger node acts as the starting point. It monitors the site’s RSS feed (https://website-name.com/feed) every five minutes and captures any new articles that appear. When a new post is detected, it passes the article title, link, and content snippet to the next node.
2. Edit Fields (Data Preparation)
The Edit Fields node structures and cleans the data from the RSS feed.
It extracts specific values from the raw feed and prepares them for AI processing.
- Title: Extracted from $json.title
- Content: Extracted from $json[‘content:encodedSnippet’]
- Link: Directly from $json.link
- Image URL: Captured through regex from the HTML <img> tag embedded in the content
This node ensures the workflow always passes clean, consistent data to the AI engine.
3. Basic LLM Chain (AI Caption Generator)
This node uses an LLM Chain connected to OpenAI GPT-3.5 turbo model to generate optimized social media captions. The node combines the article’s title and description into a structured prompt asking the model to:
Write a five-sentence summary of the job post in an engaging tone.
Add fifteen hashtags, grouped into several categories
4. Model (GPT-3.5 turbo)
The AI Model node defines the AI engine used by the previous LLM Chain.
It runs GPT-3.5-Turbo through OpenAI API connection and handles all prompt execution and response generation.
5. Merge Node (Data Consolidation)
Once the LLM chain generates the caption, the Merge node combines the structured data from the RSS feed with the AI output. Each post now has following elements.
Title
Generated description
Image
Link to the original article
Generated hashtags
6. Publishing to Facebook and Instagram
The workflow then branches into Facebook Graph API nodes. In this example I have configured to publish to 3 separate Facebook pages each with different page credentials.
Each 3 nodes have query parameters as follows.
message: Title + AI-generated caption + Link to the original article + Hashtags
url: The job image url from the RSS feed
This allows workflow to share the same post simultaneously to multiple accounts while maintaining independent access tokens.
For Instagram, posting is handled through two sequential nodes because the Instagram Graph API does not allow publishing media in a single request.
- Upload Media – This step sends the image (and optional caption) to Instagram via the Graph API. Instagram validates and processes the file, then returns a creation_id.
- Publish Media – Using that creation_id, the workflow triggers the actual publish action, which makes the post go live on the profile.
Also since Instagram does not support clickable links in captions the workflow automatically adds a “Link in Bio” note so users know where to find the external link.
Results of the POC
The workflow has been running continuously for the past 3 months without any hiccups except for a single event caused by a Graph API key expiration. Each time a new post is published on the website it now appears across 3 Facebook pages and one Instagram account within minutes, formatted with professional captions and relevant hashtags.
The posts are visually consistent, timely, and engaging. All of this achieved without any manual copy pasting, scheduling, or rewriting. The POC clearly demonstrates that social media publishing can be automated at scale using n8n with minimal maintenance.
Potential Similar Applications
This approach is not limited to this. The same architecture can be extended to other areas.
- Automatically post breaking news to large number of social media channels from the news website.
- Share new product listings, promotions or price drops across multiple social media platforms.
- Convert blog posts or press releases into shareable LinkedIn and Facebook articles.
- Instantly distribute event notices or registration updates.
- Use private feeds for team announcements across internal social channels.
By connecting n8n with Gen AI platforms, content hosting, and social APIs organizations can automate nearly every stage of their digital communication cycle.