An engineer’s side project is reshaping automated journalism by combining AI with trustworthy reporting methods. The Agentic Tribune represents an intriguing experiment in using AI to produce fact-based news at scale, addressing financial pressures on traditional media while maintaining journalistic integrity. This project offers a window into both the technical challenges of building AI content systems and their potential implications for the future of news production.
The big picture: A software engineer working in AI has created a fully automated news site that publishes original, fact-based reporting daily without human intervention.
- The Agentic Tribune produces approximately 22 articles on weekdays and 12 on weekends, each costing between $0.25-0.30 in API usage.
- The system focuses on neutral, factual reporting based on credible sources like government releases, legal rulings, and financial reports.
Behind the automation: The site runs on a seven-step AI pipeline that handles everything from finding news leads to publishing finished articles.
- The process includes scanning structured feeds for leads, scoring them based on metrics like recency and newsworthiness, conducting AI-powered research, and writing articles.
- Each article undergoes AI self-critique and revision before being formatted, tagged, and published automatically.
Technical challenges: Managing LLM behavior proved to be one of the most significant hurdles in building a reliable automated news system.
- The developer encountered issues with models misinterpreting scoring metrics, leaking details between tasks, hallucinating facts, and misrepresenting themselves as established news organizations.
- Simpler prompts often outperformed elaborate ones, and positive requirements worked better than negative restrictions.
The tech stack: The project combines straightforward web technologies with AI capabilities.
- The backend uses Flask, SQLite, cron jobs, and Python scripts with prompts generated from Jinja2 templates.
- The system is deployed on EC2 with Nginx and Gunicorn, using static HTML for the frontend and ChatGPT-generated icons.
Why this matters: The project demonstrates how AI might complement rather than replace human journalism, potentially freeing up resources for more complex reporting tasks.
- The developer acknowledges that AI news can only report on what’s already public, emphasizing there’s “no substitute for investigative reporters, war correspondents, or interviews.”
- Low-cost, trustworthy summaries of events could help redirect human journalistic talent toward reporting that requires human judgment and investigation.
How I Built an AI-Generated News Site - and What I Learned