TrendRadar

TrendRadar is an AI-powered trending topic aggregator and public opinion monitoring tool designed for the Chinese internet. It covers 11 major platforms by default (Zhihu, Weibo, Douyin, Bilibili, Cailian Press, etc.) with extensibility up to 35, plus RSS/Atom feed support. Three push modes—daily summary, current leaderboard, and incremental monitoring—serve different user needs from investors to content creators, with delivery to 9 channels including WeCom, Feishu, DingTalk, Telegram, and email.

For content filtering, beyond traditional keyword matching (with required/exclusion syntax), v6.5.0+ introduces AI intelligent filtering where users describe interests in natural language and AI auto-classifies and scores items, with automatic fallback to keyword mode on failure. Analysis capabilities include timeline tracking, persistence analysis, cross-platform comparison, and a customizable ranking algorithm (ranking weight 60%, persistence 30%, quality 10%).

Since v3.0.0, the project integrates an MCP (Model Context Protocol) server exposing 13 analysis tools (basic queries, trend analysis, sentiment analysis, etc.), enabling conversational data queries through AI clients like Claude Desktop, Cursor, and Cherry Studio. v5.2.0+ adds AI multi-language translation, and mcp-v4.0.0+ supports direct AI-generated content push to all 9 channels.

Deployment options include GitHub Actions (fork-and-run, zero server), Docker (wantcat/trendradar:latest), and local scripts. Reports are auto-generated as polished web pages via GitHub Pages (responsive, dark mode), with a web-based graphical configuration editor available.

Note: The entry URL is a fork by user bobkingdom (stuck at v3.0.4); the official repository is sansan0/TrendRadar (current v6.6.0). Information above is based on the official repo. The Docker Hub image maintainer identity and the specific AI model used for filtering are not explicitly documented.