WordPress Content Machine with Human-in-the-Loop Review - n8n Workflow

Content Agencies scaling their blog production.
Technical SEO specialists looking to automate large-scale content creation.
Developers needing advanced examples of LangChain agents integrated into an n8n workflow.
Users seeking robust n8n templates for generative AI and publishing.

Managing high-quality, research-intensive content generation at scale is complex. This powerful n8n workflow solves this by creating a structured pipeline combining AI capabilities (research, writing, image generation) with a critical human-in-the-loop (HITL) review process managed entirely through Airtable.

This robust n8n automation moves beyond simple content generation by incorporating deep external research (via the Research Tool), generating featured and chapter-specific images, calculating SEO metadata (like internal links from a sitemap via the HTTP Request n8n node), and finally publishing the fully formatted HTML to WordPress. The modular design, heavily relying on the Airtable Trigger n8n node, makes this an incredibly flexible and scalable solution for demanding content operations. This is one of the most comprehensive n8n templates available for production-level content.

The entire process is controlled by state changes in Airtable, which acts as the project management layer. This single n8n workflow handles four distinct stages:

1. Topic Initialization: The Create New Topic n8n trigger detects a new article request in Airtable. After filtering, the process sends the topic idea to the Initial Research LangChain agent. This agent uses an OpenAI Chat Model and a custom workflow-based Research Tool to gather initial data and structure the article outline. The automation then updates the Airtable record, moving it to a "Review Required" status (the human-in-the-loop step).


2. Deep Chapter Generation: When the human reviewer approves the structure, the Select Chapters Trigger n8n node fires. It dynamically retrieves internal links (using HTTP Request to fetch an XML sitemap) and passes this SEO context to the Researcher agent. This agent performs in-depth research for each chapter. The subsequent Copywriter chain takes the research and uses a Structured Output Parser to generate the final, high-quality chapter text, which is saved back to Airtable.


3. Article Assembly & Media: The Airtable Select Content Trigger starts the assembly stage. A Blog Planner agent structures the final article metadata and prompts DALL-E (via the Generate featured image n8n node) and a loop of Generate Chapter Image nodes to create all necessary visuals. Images are resized (Edit Image), uploaded to Google Drive, and then posted to the WordPress Media Library via HTTP Request.


4. Final Publishing: Once the content is assembled and approved, the Airtable Finalize Post Trigger activates. The content (converted from Markdown to HTML) and image URLs are merged. The Post on Wordpress n8n node publishes the article, and subsequent HTTP Request nodes set the featured image and excerpt. Finally, the n8n workflow updates the status in Airtable to complete the cycle.

1. Import the n8n Workflow: Copy the provided JSON and import it into your n8n instance using the 'New' -> 'Import from JSON' feature.


2. Airtable Credentials: You will need to configure multiple Airtable n8n node credentials (e.g., Airtable Chapters, Airtable Finalize Post). Ensure these credentials have access to your Content Management base and tables.


3. AI Services: Configure the OpenAI Chat Model n8n node credentials. This workflow relies heavily on robust AI access.


4. WordPress Connection: Set up the credentials for the dedicated WordPress n8n node, and ensure the necessary HTTP Request nodes (used for advanced image handling and metadata updates) are configured with your WordPress REST API access tokens.


5. External Dependencies: Configure Google Drive credentials for image backup and ensure the Research Tool (if an external workflow is referenced) is accessible and operational. This complex n8n template requires careful configuration across all connected services to run smoothly.

Airtable Trigger (Multiple Instances): These are the core control points (e.g., Create New Topic, Select Chapters Trigger, Airtable Finalize Post Trigger). They monitor the Airtable content base for status changes, making this an event-driven n8n workflow.
OpenAI Chat Model (LangChain LM): Used extensively by various agents (Initial Research, Researcher, Copywriter) to interact with advanced LLMs, generating structured outlines and final content.
Initial Research (LangChain Agent): A dedicated n8n node that uses the AI model and potentially external Research Tool workflows to produce the initial content outline.
Researcher / Copywriter (LangChain Agents): Specialized agents responsible for deep-dive content creation per chapter, ensuring the output is high-quality and factual.
Structured Output Parser (LangChain Node): Ensures AI outputs (like chapter structures or blog planner data) conform to a strict JSON schema, which is vital for the downstream logic in the n8n workflow.
HTTP Request / WordPress Nodes: Used for publishing. The Post on Wordpress n8n node handles the main posting, while multiple HTTP Request n8n nodes manage image uploads, metadata updates, and external SEO data fetching (sitemap parsing).
Code Node (Settings, Parse Chapters, etc.): Used for custom data manipulation, variable setting, and structuring inputs/outputs for the AI nodes, demonstrating flexible scripting within an n8n workflow.
Split In Batches / Split Out: These n8n node types handle complex looping structures, allowing the workflow to iterate over multiple chapters or content pieces efficiently during the AI generation phases.