Newsly is a self-hosted, AI-powered news aggregation and verification platform designed to combat misinformation and promote accountability. By leveraging a multi-agent system, Newsly scrapes news from various sources, cross-verifies the information, and generates reliable articles and podcasts. The system is designed to ensure transparency in reporting by tracking statements from politicians, crime justice statuses, and other critical societal topics.

• Eliminate fake news and misinformation by cross-referencing multiple sources.
• Provide a reliable and unbiased news aggregation system.
• Track and verify political statements and crime justice updates.
• Benchmark various open-source LLMs for performance and accuracy.
• Ensure data transparency and accountability in news reporting.

• Multi-Agent Orchestration: Autogen (AG2) for coordinating AI agents.
• LLMs (Local & Cloud):
  • Ollama (Llama 3.2-3B) for local AI processing.
  • Google Gemini AI for cloud-based model inference.
• Web Scraping:
  • Playwright & Stagehand for automated news extraction.
  • SearxNG for local search-based information retrieval.
• Retrieval-Augmented Generation (RAG):
  • Qdrant Vector Database for efficient data indexing and retrieval.
• Containerization & Deployment:
  • Docker for self-hosted deployment.
  • API integrations with cloud services.

• AI-driven web scraping and real-time news retrieval.
• Cross-verification of news across multiple sources.
• Automated news article and podcast generation.
• Self-hosted infrastructure for data privacy.
• Benchmarks open-source LLMs for accuracy and efficiency.
• Tracks and monitors public figures' statements and key societal trends.

• Optimizing AI agents for autonomous multi-tasking and orchestration.
• Enhancing RAG pipelines for real-time information verification.
• Efficient scraping using Playwright and Stagehand.
• Deploying AI-driven applications using Docker for self-hosting.
• Benchmarking various open-source LLMs for performance comparison.

Newsly is actively being developed to test and benchmark open-source LLMs and explore advancements in AI-driven news verification. Future updates will include improved accuracy, better agent collaboration, and more comprehensive tracking of political and societal trends.

If you're interested in contributing, feel free to open issues or submit pull requests. Suggestions for improving AI accuracy, integration with additional sources, and performance optimizations are welcome!

This project is open-source under the MIT License.

Stay tuned for updates and improvements as we push the boundaries of AI-driven journalism!