Steven Morlier

RSS-to-Image Summarizer

A pipeline that read the day's news from RSS feeds, summarized it, and turned each story into a Midjourney prompt for a cat-themed illustration.

  • RSS
  • Automation
  • AI

Project Overview

This experimental project combined Python scripting, RSS feeds, Google Sheets data manipulation, and AI-generated imagery to explore a novel way of visualizing daily news. By fetching current events from RSS feeds, the system summarized articles and then generated prompts for Midjourney. Although no direct API connection was established, the prepared prompts could be manually pasted into Midjourney to produce unique cat-themed images representing the news.

Key Features

Automated RSS Retrieval
Continuously fetched news articles from selected RSS feeds.
Summarized News Content
Condensed multiple articles into concise summaries for quick review.
Google Sheets Integration
Employed Python scripts to interact with Google Sheets, automatically updating and formatting prompts.
Midjourney Prompt Generation
Produced compatible text prompts that could be copied into Midjourney to generate images.

Technical Highlights

Google Colab Environment
Leveraged cloud-based Python notebooks for easy experimentation and code execution.
Sheet-Driven Workflows
Utilized Google Sheets as a dynamic interface for organizing and refining textual data before creating prompts.
Manual Image Generation Step
No direct API calls—prompt generation was automated, but image creation required manual input into Midjourney.

Impact & Outcomes

Proof-of-Concept Innovation
Showed how real-time news content could inspire AI-driven, visual representations.
Process Automation
Demonstrated the ability to streamline prompt preparation and reduce manual overhead.
Limited Adoption
Did not progress to automated publishing due to repetitive content and limited variability in visuals.

Resources

On request: code snippets, instructions, and example outputs illustrating the prompt generation process and corresponding Midjourney-generated images.