AIHOT is a self-hosted framework for building AI-curated news and industry intelligence websites. It collects material from sources such as RSS feeds, websites, JSON APIs, social accounts, WeChat, and custom scripts. Incoming items are deduplicated, prefiltered, scored twice by language models, and selected according to configurable thresholds. The system writes summaries, titles, translations, tags, and recommendations before clustering related reports into shared events. Event popularity is calculated from independent sources instead of raw article counts. It can automatically publish daily, weekly, and monthly briefings with topic pages and search. RSS, APIs, MCP, Markdown endpoints, administration tools, and Docker deployment make the framework usable by both people and AI agents.
Features
- Multi-source news collection
- AI filtering and double scoring
- Automatic summaries and translations
- Event clustering and trend ranking
- Daily, weekly, and monthly reports
- RSS, API, MCP, and agent interfaces