Horizon: what it is, what problem it solves & why it's gaining traction

What it solves

Horizon is an AI-powered news radar designed to solve the problem of information overload from scattered sources. It transforms a chaotic stream of news from platforms like Hacker News, Reddit, and RSS feeds into a curated, ranked, and enriched daily briefing, reducing noise while maintaining a human-centric approach to content discovery.

How it works

Horizon operates through a multi-step pipeline:

1. Define: The user configures sources, AI models, and delivery channels via a JSON file or an interactive setup wizard.
2. Fetch: It concurrently pulls content from configured sources (RSS, Hacker News, Reddit, Telegram, Twitter/X, GitHub, and OpenBB).
3. Deduplicate: It merges identical stories found across different platforms to avoid repetition.
4. Score & Filter: An AI model (such as GPT, Claude, or Gemini) ranks each item from 0-10 and filters out those below a specific threshold.
5. Enrich: For high-scoring items, the system performs web research to add background context and collects community discussions/comments.
6. Summarize: It generates a structured Markdown briefing with summaries and tags.
7. Deliver: The final report is published to GitHub Pages, sent via email, or pushed to webhooks (Slack, Discord, Feishu, etc.).

Who it’s for

It is for individuals who want a personalized, automated news digest of technical or niche interests without manually browsing multiple platforms daily.

Highlights

• Multi-Source Integration: Supports a wide array of sources including social media, developer platforms, and financial news.
• AI-Driven Ranking: Uses various LLMs to score and filter content based on user preferences.
• Flexible Delivery: Can be deployed as a GitHub Actions cron job to automate a daily website or newsletter.
• Contextual Enrichment: Doesn't just summarize; it adds web-researched background and community sentiment.
• MCP Server Support: Exposes the pipeline as tools for AI assistants.