An AI news filter that will find you on any messaging app

What it does

The tool pulls hot news from multiple platforms and RSS feeds, filters them by keywords, then runs an LLM over the results to translate, summarize, and generate briefs. The survivors are pushed to your device through a sprawling list of channels—WeChat, Feishu, DingTalk, Telegram, Slack, email, ntfy, Bark, and generic webhooks. It is designed to run containerized with data kept locally or in the cloud.

The interesting bit

The notification channel coverage is almost aggressive: if an app can buzz your pocket, this tool probably supports it. More technically, it exposes an MCP interface so external AI clients can query aggregated data for sentiment analysis and trend prediction rather than treating the system as a passive black box.

Key highlights

• Aggregates multi-platform hot news and custom RSS feeds, with keyword-based filtering.
• Uses LLMs to filter, translate, and generate analysis briefs before pushing to mobile.
• Supports a wide array of notification channels including WeChat, Feishu, DingTalk, Telegram, Slack, email, ntfy, Bark, and webhooks.
• Exposes an MCP interface for external AI tools to run dialogue analysis and trend prediction on aggregated data.
• Runs via Docker with support for local or cloud data storage.

Caveats

• Core trend data is sourced entirely from the third-party newsnow API, and the README explicitly warns containerized deployments to throttle requests and not hammer the service.
• The visible documentation is heavy on badges and donation appeals, so expect to scroll past sponsor walls to find technical details.

Verdict

A solid fit if you want a self-hosted, AI-gated news firehose with maximum notification reach. Skip it if you need deep primary data sourcing—the actual trend aggregation is outsourced to newsnow.