Full Production System Project – Build a Complete AI + SEO + Social + IndexNow Automation Pipeline

This capstone project builds a complete, production-grade content automation pipeline that covers the entire lifecycle: from topic selection to Google indexing. It combines every skill from the previous five modules.

What the system does:

• Reads pending topics from an Airtable content calendar
• Uses Claude to generate a 1,200-word SEO-optimized article
• Uses GPT-4o to generate the SEO title, meta description, and social media posts (4 platforms)
• Publishes the article to WordPress via REST API
• Posts to Facebook Page, LinkedIn, Twitter/X, and Telegram Channel
• Submits the new URL to IndexNow (Bing, Yandex instant indexing)
• Logs everything to MySQL + sends a completion report to Slack
• Updates Airtable record to "published" with the live URL

Total time from trigger to live article: under 45 seconds.

The content generation layer uses three AI calls in sequence, each feeding the next.

Call 1 — Claude: Article Body

POST https://api.anthropic.com/v1/messages
{
  "model": "claude-sonnet-4-6",
  "max_tokens": 3000,
  "system": "You are an expert SEO content writer. Write informative, factually accurate articles. Include one H2 per major section. Write for a general audience. Do not use AI clichés like 'dive into' or 'delve'.",
  "messages": [{
    "role": "user",
    "content": "Write a 1,200-word article about: {{ $json.topic }}\nTarget keyword: {{ $json.primary_keyword }}\nAudience: {{ $json.audience }}\nTone: {{ $json.tone }}"
  }]
}

Call 2 — GPT-4o: SEO Metadata

Generate the following for this article topic: {{ $json.topic }}
Primary keyword: {{ $json.primary_keyword }}
Article excerpt (first 200 words): {{ $json.article_excerpt }}

Return ONLY valid JSON with these exact keys:
{
  "seo_title": "...",
  "meta_description": "...",
  "slug": "...",
  "tags": ["tag1", "tag2", "tag3"],
  "focus_keyword": "..."
}

Call 3 — GPT-4o: Social Media Posts

Write 4 social media posts for this article:
Title: {{ $json.seo_title }}
Topic: {{ $json.topic }}
URL: (will be added after WordPress publish)

Return JSON:
{
  "twitter": "Max 280 chars, punchy, include 2 hashtags",
  "linkedin": "200-300 chars, professional tone, include URL placeholder [URL]",
  "facebook": "150-200 chars, conversational, include URL placeholder [URL]",
  "telegram": "150 chars, direct, with emoji, include URL placeholder [URL]"
}

After all three AI calls complete (run in sequence — Call 2 and 3 can be parallel), publish to WordPress via the REST API.

// HTTP Request: POST to WordPress
POST https://yourdomain.com/wp-json/wp/v2/posts
Auth: Basic Auth (Application Password)
Body:
{
  "title": "{{ $node["GPT-SEO"].json.seo_title }}",
  "content": "{{ $node["Claude-Article"].json.content[0].text }}",
  "status": "publish",
  "slug": "{{ $node["GPT-SEO"].json.slug }}",
  "excerpt": "{{ $node["GPT-SEO"].json.meta_description }}",
  "tags": [tag IDs from previous step],
  "meta": {
    "_yoast_wpseo_title": "{{ $node["GPT-SEO"].json.seo_title }}",
    "_yoast_wpseo_metadesc": "{{ $node["GPT-SEO"].json.meta_description }}",
    "_yoast_wpseo_focuskw": "{{ $node["GPT-SEO"].json.focus_keyword }}"
  }
}

WordPress returns the published post including its link field — the live URL. Extract it with a Set node: live_url: {{ $json.link }}. All subsequent nodes use this URL.

With the live URL available, replace the [URL] placeholder in all social posts and publish simultaneously to all four platforms.

Replace URL placeholders in a Code node:

const liveUrl = $node["WordPress-Publish"].json.link;
const socialPosts = $node["GPT-Social"].json;
return [{
  twitter:  socialPosts.twitter.replace("[URL]", liveUrl),
  linkedin: socialPosts.linkedin.replace("[URL]", liveUrl),
  facebook: socialPosts.facebook.replace("[URL]", liveUrl),
  telegram: socialPosts.telegram.replace("[URL]", liveUrl),
  live_url: liveUrl
}];

Facebook Page Post: POST to https://graph.facebook.com/v21.0/PAGE_ID/feed with message and link parameters using your Page Access Token.

LinkedIn Post: POST to https://api.linkedin.com/v2/ugcPosts with your LinkedIn OAuth2 token.

Twitter/X: POST to https://api.twitter.com/2/tweets with OAuth1.0a credentials.

Telegram Channel: POST to https://api.telegram.org/bot{TOKEN}/sendMessage with chat_id = your channel username.

IndexNow:

POST https://api.indexnow.org/indexnow
{
  "host": "yourdomain.com",
  "key": "{{ your-indexnow-key }}",
  "keyLocation": "https://yourdomain.com/{{ your-indexnow-key }}.txt",
  "urlList": ["{{ $json.live_url }}"]
}

Run all five parallel HTTP requests simultaneously, then merge with a "Wait All" Merge node.

After all publishing actions complete, run three final operations in parallel:

1. MySQL Log Entry:

INSERT INTO published_articles (
  topic, slug, live_url, word_count, seo_title,
  airtable_record_id, published_at, workflow_execution_id,
  social_facebook, social_linkedin, social_twitter, social_telegram
) VALUES (...);

2. Airtable Update: Update the original content calendar record to status "published", add the live_url, and set published_at timestamp. Use the PATCH endpoint with the record ID stored at the start of the workflow.

3. Slack Completion Report:

// Slack message format
*Published:* <{{ live_url }}|{{ seo_title }}>
*Topic:* {{ topic }}
*Word Count:* {{ word_count }}
*Social:* Facebook ✓ | LinkedIn ✓ | Twitter/X ✓ | Telegram ✓
*IndexNow:* Submitted ✓
*Total Time:* {{ elapsed_seconds }}s

Merge all three parallel operations, then send the Slack report. The entire pipeline — from Airtable trigger to Slack confirmation — runs in under 45 seconds for a 1,200-word article.