TrendRadar

TrendRadar exposes an MCP (Model Context Protocol) server that enables AI clients to connect, query, and analyze hot news data from multiple sources. It centralizes data collection, filtering, and push notifications, while offering optional AI-assisted analysis and multi-client integration for deep insights and automation.

How to use

You connect an MCP client to TrendRadar to access hot news data and analyses. You can run in several modes: fetch and push daily summaries, monitor the current top stories, or watch for only new items as they appear. You can also use AI-enabled analysis clients to perform natural-language queries and topic trends against locally stored data.

To start using, establish an MCP connection using either the remote HTTP endpoint or a local STDIO server. The HTTP method points to a live URL you provide, while the STDIO method runs a local process that exposes the MCP endpoint through standard input/output. Pick the approach that best fits your environment and security needs.

How to install

Prerequisites you need before installation: Docker and Docker Compose for containerized deployment, or Python and a compatible runtime if you choose to run locally. Ensure you have a stable internet connection and a persistent data directory for logs and reports.

# Option A: Quick Docker deployment

# Run this in any directory you want TrendRadar to manage config and output

docker run -d --name trend-radar \
  -v ./config:/app/config:ro \
  -v ./output:/app/output \
  -e FEISHU_WEBHOOK_URL="你的飞书webhook" \
  -e DINGTALK_WEBHOOK_URL="你的钉钉webhook" \
  -e WEWORK_WEBHOOK_URL="你的企业微信webhook" \
  -e TELEGRAM_BOT_TOKEN="你的telegram_bot_token" \
  -e TELEGRAM_CHAT_ID="你的telegram_chat_id" \
  -e EMAIL_FROM="你的发件邮箱" \
  -e EMAIL_PASSWORD="你的邮箱密码或授权码" \
  -e EMAIL_TO="收件人邮箱" \
  -e CRON_SCHEDULE="*/30 * * * *" \
  -e RUN_MODE="cron" \
  -e IMMEDIATE_RUN="true" \
  wantcat/trendradar:latest

# Option B: Docker Compose (recommended)

# 1) Create project structure
mkdir -p trendradar/{config,docker}
cd trendradar

# 2) Download templates
wget https://raw.githubusercontent.com/sansan0/TrendRadar/master/config/config.yaml -P config/
wget https://raw.githubusercontent.com/sansan0/TrendRadar/master/config/frequency_words.txt -P config/
wget https://raw.githubusercontent.com/sansan0/TrendRadar/master/docker/.env
wget https://raw.githubusercontent.com/sansan0/TrendRadar/master/docker/docker-compose.yml

# 3) Start the service
docker-compose pull
docker-compose up -d

# 4) Check status
docker ps | grep trend-radar

docker logs -f trend-radar

Configuration and security notes

MCP connections are configured via the provided config templates and optional environment variables. When you use HTTP mode, you expose a remote MCP endpoint. When you use STDIO mode, you run a local process that serves MCP over standard I/O. Always protect webhook URLs, tokens, and API keys. Do not publish secrets in public repositories.

Key connection options shown in examples:

• HTTP mode example URL: http://localhost:3333/mcp
• STDIO mode example command: uv --directory /path/to/TrendRadar run python -m mcp_server.server If you are using HTTP mode, the URL must be provided for the connection to work.

Notes on MCP clients and tools

The MCP client integration supports a range of tools to interact with TrendRadar. Common capabilities include querying latest news, retrieving by date, listing trending topics, and performing advanced analyses like sentiment and topic trends. You can connect multiple clients for different workflows, including STDIO-based local servers and HTTP-based remote servers.

Available tools described in the MCP context include: get_latest_news, get_news_by_date, get_trending_topics, search_news, search_related_news_history, analyze_topic_trend, analyze_data_insights, analyze_sentiment, find_similar_news, generate_summary_report, get_current_config, get_system_status, trigger_crawl.

Troubleshooting and tips

If the MCP service cannot start, verify port availability, check dependencies, and review container logs. If a client cannot connect, ensure the correct mode (http vs stdio) and verify the provided URL or directory paths. For issues with AI analysis integration, ensure local data exists in the output directory and that the MCP server is reachable by the client.

Available tools

get_latest_news

Query the most recent news entries available in the local data store.

get_news_by_date

Retrieve news entries for a specific date range.

get_trending_topics

Fetch currently trending topics across monitored platforms.

search_news

Perform a general search across stored news data.

search_related_news_history

Find news items related to a historical topic.

analyze_topic_trend

Analyze how a topic's popularity changes over time.

analyze_data_insights

Derive insights by comparing data across platforms.

analyze_sentiment

Assess sentiment of news items.

find_similar_news

Identify news items similar to a given article.

generate_summary_report

Create a concise summary report from the data.

get_current_config

Return the current MCP server configuration.

get_system_status

Provide the status of the MCP service and components.

trigger_crawl

Manually trigger a data crawl to refresh content.