Verified project record

Thysrael/Horizon

horizon is a personal AI-powered news radar that fetches, deduplicates, scores, filters, and enriches stories from various sources to produce bilingual daily briefings. It consolidates updates from platforms including Hacker News, RSS, Reddit, Telegram, Twitter/X, and GitHub into a single pipeline with English and Chinese output.

Project overview

horizon addresses information overload by combining multi-source tracking, story deduplication, AI-driven scoring, background enrichment, and community comment summarization into one pipeline that generates bilingual briefings through multiple delivery channels.

Project type: MCP
Deployment: Refer to project documentation
License: MIT

Best for

Developers and researchers who want a self-hosted, customizable news radar that consolidates multiple platforms into bilingual daily briefings with AI-driven scoring and filtering.

Key capabilities

Tracks Hacker News, RSS, Reddit, Telegram, Twitter/X, GitHub releases or user activity, and OpenBB financial news watchlists in one pipeline.
Merges the same story across platforms before it reaches the briefing.
Analyzes each item with a stable processing profile and applies user-defined filter thresholds.
Adds web-researched context for unfamiliar concepts, companies, projects, and technical terms.
Collects and summarizes community comments from supported sources like Hacker News and Reddit.
Generates English and Chinese daily briefings from the same source set.
Publishes generated Markdown as a GitHub Pages daily briefing site.
Runs a self-hosted SMTP/IMAP newsletter with automatic subscribe and unsubscribe handling.

Limitations and risks

The Twitter extra requires a Playwright browser and system packages that are not installed by the current Dockerfile.

Getting started

Clone the repository, install dependencies with uv or pip, configure config.json and .env with API keys and data sources, then run the horizon CLI. An interactive setup wizard can generate a personalized source configuration based on user interests. An external LLM API, such as OpenAI or Gemini, must be configured for AI analysis.

Evidence and sources

Open original GitHub project