News Agents – An open – source terminal – based news aggregation and summarization system

What is News Agents？

News Agents is a terminal-based news aggregation and summarization system. Built on the Amazon Q CLI as its agent framework, it uses the Model Context Protocol (MCP) to parse RSS news feeds and leverages tmux for terminal window splitting and multitask monitoring. The system fetches articles from various news sources such as Hacker News, TechCrunch, and The Wall Street Journal, distributes them to multiple sub-agents for parallel processing, and generates concise summaries. These summaries are then consolidated into a main-summary.md file. The entire workflow is executed within the terminal, showcasing the powerful collaboration between multiple agents and providing users with an efficient, personalized news reading experience.

Key Features of News Agents

• News Aggregation: Collects RSS feeds from multiple well-known sources such as Hacker News, TechCrunch, and The Wall Street Journal.

• Multi-Agent Collaboration: Utilizes a main agent and multiple sub-agents to process news feeds in parallel, enhancing efficiency.

• News Summarization: Each sub-agent parses the news content and generates concise summaries.

• Terminal Visualization: Enables real-time monitoring of each agent’s processing status via terminal panes.

• Consolidated Output: Aggregates all summaries generated by sub-agents into a single main summary file (main-summary.md) for quick browsing.

Technical Architecture of News Agents

• Amazon Q CLI: Serves as the agent framework for managing and orchestrating the main and sub-agents. It provides tools and interfaces to support multi-agent collaboration and task allocation.

• Model Context Protocol (MCP): Parses RSS news feeds to extract headlines, links, and content. Offers a standardized interface to facilitate processing across different sources.

• tmux: Splits the terminal into multiple panes to display the runtime status of both the main agent and sub-agents, enabling real-time visualization and multitasking.

• Parallel Processing: Divides the list of news sources into segments, each processed by a separate sub-agent. With tmux, each agent’s activity is clearly visible in its own pane, improving efficiency and transparency.

Project Links for News Agents

• Project Website: https://eugeneyan.com/writing/news-agents/

• GitHub Repository: https://github.com/eugeneyan/news-agents

Application Scenarios of News Agents

• Personal News Reading: Enables users to quickly access summarized content from multiple news sources directly in the terminal, saving browsing time.

• Information Filtering: Helps users quickly identify news items of interest and avoid information overload.

• Developer Testing: Provides an experimental framework for developers to test multi-agent collaboration and news processing functionalities.

• News Research: Allows researchers to aggregate and analyze content from various sources efficiently.

• Terminal Enthusiasts: Offers a fun and practical tool for users who prefer working in the terminal to read and manage news content.