Automate WordPress Content Publishing Using n8n & GPT

This guide walks you through building a serverless AI-powered publishing system that fetches AI news, rewrites it with GPT, and posts it directly to your WordPress site — without touching sensitive data like passwords or API keys in this tutorial.

✅ Overview

The system:

Fetches trending news from an AI news API
Filters and parses relevant articles
Scrapes full HTML content
Cleans, extracts, and rewrites with GPT
Formats final content
Publishes directly to WordPress
Notifies Slack (optional)

🧩 n8n Workflow Breakdown

🔹 1. Schedule Trigger

Runs your automation at regular intervals (e.g., every 6 hours or daily).

Node: Schedule Trigger

Configure for your desired frequency (e.g., every 6 hours)

🔹 2. Fetch AI News

Uses a news API (e.g., NewsAPI or Bing News Search) to retrieve the latest AI-related articles.

Node: HTTP Request

Method: GET
Endpoint: Example – https://newsapi.org/v2/everything?q=artificial+intelligence
Return only article fields like title, url, source, publishedAt, description.

🔹 3. Extract `articles[]` Array

Extracts the articles array from the JSON response.

Node: Set or Function

Outputs: array of individual articles to process

🔹 4. Loop Over Items

Iterates through each article to scrape full content.

Node: Loop Over Items

🔹 5. HTTP Request

Fetches the full HTML of the article page using its url.

Node: HTTP Request

Method: GET
URL: {{$json["url"]}}
Response Format: String

🔹 6. HTML Extract

Extracts the article body from the HTML using a CSS selector.

Node: HTML Extract

Target: article, .content, or .post-body (adjust per website)
Extract title, meta description, and content sections

🔹 7. Pre-clean & Decide What Needs AI

Function node strips HTML, removes boilerplate, and checks if GPT should rewrite title or description.

Key Tasks:

Strip tags
Remove nav/menu content
Flag missing or generic titles/descriptions

🔹 8. Message a Model (GPT)

Sends content to OpenAI (or Claude) to rewrite.

Prompt:

plaintextCopyEditYou are a professional journalist. Rewrite the article below. Return only this JSON:
{
  "cleanedContent": "...",
  "suggestedTitle": "...",
  "summary": "..."
}
Title: {{ $json.existingTitle }}
Description: {{ $json.existingDescription }}
Raw Content: {{ $json.content }}

Node: OpenAI or Claude Message node
Temperature: 0.7
Max tokens: 1024+

🔹 9. Format Article

Formats GPT output into proper JSON structure for WordPress:

title
content
excerpt
tags (optional)
categories (optional)

🔹 10. Post to WordPress

Posts the article using the WordPress REST API.

Node: HTTP Request

Method: POST
URL: https://xtreamsolution.net/wp-json/wp/v2/posts
Auth: Use Basic Auth with an Application Password
Body:

jsonCopyEdit{
  "title": "{{$json.title}}",
  "content": "{{$json.content}}",
  "excerpt": "{{$json.excerpt}}",
  "status": "publish",
  "categories": [123],
  "tags": ["AI", "automation"]
}

🔒 Don’t hardcode API keys or passwords directly in nodes. Use n8n credentials or environment variables.

🔹 11. Notify via Slack (Optional)

Sends a message to a Slack channel confirming the post.

jsonCopyEditNew article posted: *{{$json.title}}*
<https://xtreamsolution.net>

✅ Deployment Tips

Test each section independently before connecting everything.
Inspect CSS on source websites to find reliable article containers (<article>, .post-body, .cna-article-content, etc.).
Rate limit GPT calls if needed via delay or batching.
Use categories & tags for better SEO on WordPress.

🧠 Optional Enhancements

Add a duplicate-check node to prevent reposting
Generate slug from title
Schedule GPT summaries for newsletter digests
Add Google Sheets or Airtable backup

Automate WordPress Content Publishing Using n8n & GPT