What are the technical requirements to run Horizon?

Horizon requires Python 3.11+ and access to the Horizon repository. It also needs a valid configuration file and can optionally utilize API keys for services like OpenAI or GitHub.

How does Horizon process news content?

Horizon executes a full content processing pipeline, including fetching, AI-powered scoring for relevance, filtering and de-duplication, enriching high-priority content, and generating summaries.

Horizon is a backend server for the Horizon system, providing automated stages to process, analyze, and summarize news content, from fetching to final summary generation.

What kind of summaries does Horizon generate?

Horizon generates concise Markdown summaries from processed and enriched news content, making it easy to consume key information.

Does Horizon use Artificial Intelligence?

Yes, Horizon leverages AI specifically to score news content, determining its relevance and helping to prioritize information during the processing stages.

Horizon

Name: Horizon
Author: henry-insomniac

•

Provides a backend server for the Horizon system to process, analyze, and summarize news content through a series of automated stages.

Horizon MCP is an MCP Server designed for the Horizon news processing system. It leverages Horizon's native Python implementation to expose a suite of tools for managing the news lifecycle. From fetching and de-duplicating raw news items to scoring with AI, filtering by relevance, enriching with background context, and generating Markdown summaries, Horizon MCP orchestrates the entire pipeline. It ensures consistent core logic, stage re-entrancy, and minimal external side effects, making it a robust and reliable component for content curation.

主な機能

01Filter and de-duplicate news items based on scores and topics

02Enrich high-priority content with additional background information

03Execute full content processing pipelines for news (fetching, scoring, filtering, enriching, summarizing)

04Score news content using AI for relevance

05Generate Markdown summaries from processed content

063 GitHub stars

ユースケース

01Generating structured summaries from filtered and enriched news

02Automated news content processing and analysis workflows

03Integrating content curation functionalities into MCP client applications