Case StudyDocker VPS + WordPress - Production

Jobs Publishing Agent

A high-volume production n8n automation that discovers job posts, extracts links with custom JavaScript, prevents duplicates through Airtable, uses an AI agent with structured output parsing to rewrite Arabic SEO job content, publishes to WordPress, uploads featured images, sends Telegram notifications, and forwards structured data to an external platform webhook.

Jobs Publishing Agent

Core Tech Stack

  • n8n
  • OpenAI
  • WordPress
  • Airtable
  • Telegram

Performance

  • Runs automatically on a short schedule without manual operation
  • Duplicate checks happen before costly AI generation
  • Canonical URL hashing keeps lookup data compact
  • Single workflow handles discovery, enrichment, publishing, and notifications
  • Production history shows high reliability across more than 1300 reviewed executions

Security

  • Credentials are managed inside n8n and kept out of public project data
  • Duplicate gate prevents accidental content spam
  • Structured parser reduces malformed AI data risk
  • External webhook forwarding uses server-side authentication
  • Private workflow exports are sanitized before public portfolio usage

Key Features

Scheduled job discovery workflow running in production

Custom JavaScript parser for extracting job links and titles

URL normalization and canonical key generation

Duplicate detection using Airtable search records

AI Agent for factual extraction and Arabic SEO rewriting

Structured Output Parser for predictable job data

Automatic sector, city, and employment type categorization

WordPress post creation with generated content

Image download, media upload, and featured image assignment

Telegram notification generation using AI

JobPosting JSON-LD schema generation inside content

Reviewed production history of 1300+ executions with a very high success rate

Project Architecture

Schedule Trigger
├── Fetch News List scrapes job source
├── Code node extracts the first job link
├── Titles node normalizes data
├── Airtable Search checks duplicate records
├── IF gate allows only new jobs
├── Airtable Create stores unique key
├── AI Agent extracts and rewrites job content
├── Structured Output Parser validates output shape
├── Category Code detects WordPress category
├── Download and Upload Image handle media
├── WordPress creates and updates the post
├── AI creates Telegram caption
└── Telegram sends the published update

Engineering Strategies

  • Used custom parsing instead of fragile manual copy/paste workflows
  • Stored unique hashes in Airtable to prevent duplicate WordPress publishing
  • Constrained AI output with a manual schema for reliable automation
  • Kept factual fields intact while rewriting content into original Arabic copy
  • Added SEO-focused rules and JobPosting JSON-LD generation for search visibility
  • Split publishing into content creation, media upload, featured image update, and notification steps
  • Added external webhook forwarding to reuse the generated structured job data elsewhere

Visual Interface

Jobs Publishing Agent screenshot 1
Jobs Publishing Agent screenshot 2

Interested in this project?

I'm available for new opportunities and collaborations. Let's build something exceptional together.

Get In Touch
Mostafa Abdellraheem | Full-Stack Engineer & AI Automation Specialist