AI News Dashboard

A modern web application that aggregates and displays AI-related news articles from three major sources: YouTube, Anthropic, and OpenAI. The application features a visually attractive dashboard with an intelligent backend pipeline powered by AI agents.

Visit

https://blog-agent-alpha.vercel.app/

Overview

The AI News Dashboard provides users with a curated feed of the latest AI news, organized by source category. Users can trigger backend scrapers to fetch fresh content and optionally receive digest emails.

Application Workflow

The application executes a multi-stage AI-powered pipeline when triggered:

1. Scraper Stage (scrapper)

   • Fetches raw articles from YouTube, Anthropic, and OpenAI sources
   • Collects metadata: title, summary, URL, publication date

2. Anthropic Stage (anthropic)

   • Uses Google AI to analyze and enhance article summaries
   • Extracts key insights and relevance scores
   • Ensures content quality and consistency

3. YouTube Stage (youtube)

   • Processes video content from YouTube channels
   • Generates video descriptions and relevance metadata

4. Digest Stage (digest)

   • Compiles processed articles into a structured digest
   • Organizes content by category and relevance

5. Email Stage (email)

   • Optionally sends the compiled digest to the provided email address
   • Formats content for readability

Each stage displays real-time progress with status indicators (loading, success, error).

AI Agents

The application leverages **Google ADK ** for intelligent content processing:

Integration Points

• Content Analysis: Analyzes raw article content to generate quality summaries
• Information Extraction: Extracts key concepts, entities, and relevance from AI news
• Quality Assurance: Ensures consistency and accuracy across different sources
• Digest Compilation: Intelligently organizes and formats digests for readability

AI Agent Architecture

The backend implements a modular agent-based pipeline where each stage (scraper, Anthropic, YouTube, digest, email) operates as an independent agent that can:

• Execute asynchronously
• Report success/failure status
• Pass structured data to the next stage
• Handle errors gracefully

This design allows for:

• Resilience: Failure in one stage doesn't cascade without proper error handling
• Maintainability: Each agent has a single responsibility
• Real-time Feedback: Progress updates visible to users as each stage completes

Technology Stack

Frontend:

• Next.js 20 with React 19

Backend:

• Flask for API endpoints
• Google ADK integration

Infrastructure:

• Full-stack JavaScript/Python hybrid application

Features

✨ Dashboard - Visual interface with category-based design

📱 Responsive Design - Works seamlessly on mobile, tablet, and desktop

🤖 AI-Powered Processing - Google analyzes and enhances article content

📧 Email Digests - Optional email delivery of curated news

🎯 Time Range Filtering - Control how recent articles should be

🔄 Real-time Progress - Visual pipeline status during background processing

🔗 Direct Article Links - One-click access to full articles