Super RAG Fact-Based Article Generator from Knowledge Base - n8n Workflow

• Content teams needing to rapidly generate high-quality, verified articles.


• Companies relying on specific, proprietary internal knowledge (RAG) for external communications.


• Advanced n8n users looking for complex AI automation templates.


• Marketers who need to scale content production while maintaining factual accuracy.

This automation template represents a powerful solution for scaling content creation while strictly enforcing factual accuracy. Standard AI writing often struggles with hallucinations; this n8n workflow eliminates that risk by grounding the entire article generation process in a verified knowledge source (via Super Assistant).

The process begins with an easy-to-use n8n trigger form where the user submits a topic and guidelines. The core innovation of this n8n workflow is the 'plan, research, write' sequence: an LLM first breaks the topic into precise sub-questions, which are then individually answered by querying your RAG knowledge base. Finally, a second, more powerful LLM writes the final article based only on the verified research data, ensuring source links are integrated seamlessly. This complex flow demonstrates the advanced capabilities available through pre-built n8n templates.

This highly structured n8n workflow operates in five distinct phases:

1. Trigger and Planning (The n8n trigger): The automation starts with the New article form n8n trigger, capturing the desired article title and guidelines. The New content - generate research questions n8n node, powered by GPT-5, analyzes the topic and decomposes it into 5–8 non-overlapping research questions in a strict JSON format.


2. Preparation and Split: The Structured Output Parser ensures the questions are correctly formatted, and the Split Out Questions n8n node converts the array into individual items for parallel processing.


3. RAG Research Loop: The Loop Over Questions n8n node initiates a batch loop. Within the loop, the Query Super Assistant n8n node sends each individual question to the Super RAG service (Super.work) to retrieve factual answers and relevant sources from your connected knowledge base. The Format question and answer n8n node then cleans and standardizes the output, especially replacing source placeholders with clear markdown links.


4. Aggregation: Once the research loop completes, the Aggregate research content n8n node collects all verified Q&A data, forming the final research brief.


5. Final Synthesis: The New content - Generate the AI output n8n node (using GPT 5 chat) receives the original prompt and the aggregated research. It is instructed to act as a writer, producing a full article strictly based on the provided facts. The result is captured in the final Article result n8n node, completing the automated n8n workflow.

To deploy this comprehensive n8n workflow, follow these steps:

1. Import the n8n template: Copy the provided JSON code and paste it into your n8n instance using the 'New' -> 'Import from JSON' function.


2. Configure LLM Credentials: Connect your preferred AI provider (e.g., OpenAI) to the two Langchain LLM nodes (New content - generate research questions and New content - Generate the AI output). Ensure the credentials are set up for access to advanced GPT models (like GPT-5 mini and chat models).


3. Set up Super RAG Connection: Navigate to the Query Super Assistant n8n node. This is a crucial step for the n8n workflow's RAG component.

1. Activate: Save the n8n workflow and click 'Activate' (or 'Run' for testing) to enable the form n8n trigger.

New article form (n8n trigger): The starting point for this n8n workflow. It collects Article title and Article guidelines from a user via a webhook-triggered form.
New content - generate research questions (Langchain LLM n8n node): Uses a GPT-5 model to break down the complex article topic into a list of specific, answerable research questions, enforcing structure via a JSON output schema.
Query Super Assistant (HTTP Request n8n node): The RAG integration point. This n8n node sends individual research questions to the external Super API endpoint, leveraging your proprietary documents for fact retrieval.
Loop Over Questions (Split In Batches n8n node) & Aggregate research content (Aggregate n8n node): These n8n nodes handle the iterative research process. The Split In Batches n8n node manages the loop structure, while the Aggregate n8n node collects all individual research results into one comprehensive data set for the writer LLM.

• New content - Generate the AI output (Langchain LLM n8n node): The final synthesis step. This powerful n8n node takes the combined research and the original guidelines to draft the complete, source-linked article, adhering to the requested format and style.