About AI Daily News

AI Daily News is an AI news platform that gets news through APIs, then filters the most important stories, extracts key points from each using LLMs, and publishes them with original sources – so you can read the full context if you want to explore further.

How It Works

1. Research – Fetches articles from multiple news APIs (APITube, NewsData.io) using AI-related keywords.
2. Deduplication – Groups similar articles using semantic embeddings and cosine similarity (so you see only one story per event).
3. Ranking – Scores each story on impact, novelty, and audience interest using a large language model (Groq).
4. Summarisation – Generates a concise summary, "why it matters", and 3 key points for each story.
5. Illustration – Creates an editorial image prompt and generates a unique image using Cloudflare Workers AI (with fallback to Pollinations.ai).
6. Validation & Publishing – Validates all content, saves to a Supabase database, and sends each story as a separate Telegram message.

Tech Stack

• Orchestration – LangGraph (Python agentic workflow)
• Backend API – FastAPI (Python) – planned
• Frontend – Next.js 14, TypeScript, Tailwind CSS
• Database – Supabase (PostgreSQL + Storage)
• LLM – Groq (openai/gpt-oss-120b)
• Image Generation – Cloudflare Workers AI + Pollinations.ai (fallback)
• Messaging – Telegram Bot API

Why This Project?

This project demonstrates production‑grade AI engineering skills, including agentic orchestration, semantic deduplication, LLM‑based ranking, image generation, multi‑source research, and end‑to‑end automation – all built with clean, modular code following SOLID principles.