Autonomous Content Publishing Pipeline — User-Guided Retrieval, AI Generation, and Storage-First Publishing
From topic input to published content, grounded by user-defined sources
This system automates the creation and publishing of blog content. A user submits a topic through a native n8n form, defines the sources to be used for retrieval, and the workflow generates a complete post and featured image. All content is stored in a database and rendered by a frontend application.
Context & Problem
Producing consistent, relevant content requires research, writing, and media creation. The goal of this project was to design a repeatable pipeline that could transform a simple topic input into a complete, publishable post — while keeping content grounded in sources explicitly chosen by the user.
What the System Does
- •Accepts a topic and source list via a native n8n form
- •Retrieves contextual information from user-specified sources
- •Generates the full article (title, structure, and body)
- •Generates a featured image for the post
- •Stores content and metadata in Supabase
- •Exposes published content to a frontend that reads directly from the database
The pipeline is designed as an end-to-end system, not a one-off content generator.
High-Level Architecture
n8n orchestrates the full lifecycle. AI is responsible for content and image generation, while storage and publishing are handled independently by the frontend.
Retrieval & Grounding Model
Instead of relying on generic context, this pipeline retrieves information only from sources explicitly defined by the user. Retrieved content is injected into the generation prompt, anchoring the output to known references.
Engineering Highlights
User-controlled retrieval
The user defines which sources are used, making content grounding explicit and transparent.
Separation of concerns
AI handles generation, while n8n manages orchestration and Supabase handles persistence.
Storage-first publishing
Posts are stored before being exposed to the frontend, decoupling generation from rendering.
Deterministic pipeline execution
Each run follows the same execution path, making the system predictable and repeatable.
Extensibility
Additional steps such as scheduling, categorization, or enrichment can be added without redesigning the pipeline.
System Maturity
This pipeline is designed to run continuously without manual intervention. It produces consistent outputs while remaining flexible enough to evolve as publishing needs change.
Tech Stack
"This project demonstrates how to build a grounded, storage-first content pipeline that turns a user-defined topic and sources into a complete, publishable asset."