MCP Complete Query Guide
Below is the complete MCP tool usage Q&A, helping you learn how to use various analysis features through natural language conversation with the AI. Each category contains detailed question examples, tool behavior explanations, and adjustable parameters.
You don't need to memorize tool names or parameters. Just ask in everyday language, and the AI assistant will automatically pick the right tool.
Default Settings Explanation
The following optimization strategies are adopted by default, mainly to save AI token consumption:
- Result Limit: Default returns 50 news items. Say "return top 10" or "give me 100 items" to adjust.
- Time Range: Default queries today's data. Say "query yesterday", "last week" or "Jan 1 to 7" to specify.
- URL Links: Default no links (saves ~160 tokens/item). Say "need links" or "include URLs" to enable.
- Keyword List: Default does not use frequency_words.txt to filter news. Only used when calling "trending topics" tool.
The choice of AI model directly affects tool call effectiveness. The smarter the AI, the more accurate the calls. When you remove restrictions (e.g., from querying today to querying a week), token consumption may multiply.
This project provides a dedicated date parsing tool that can accurately parse natural language date expressions like "last 7 days", "this week", ensuring all AI models get consistent date ranges.
Basic Queries (Q1-Q3)
Q1: How to view the latest news?
You can ask like this:
- "Show me the latest news"
- "Query today's trending news"
- "Get the latest 10 news from Zhihu and Weibo"
- "View latest news, need links included"
Tool returns the latest 50 news items from all platforms by default, without URL links.
AI usually auto-summarizes, only showing partial news (like TOP 10-20 items). If you want to see all 50, need to explicitly request: "show all news" or "list all 50 items completely". This is the AI model's natural behavior, not a tool limitation.
Adjustable Specify platform, adjust quantity, include links, request full display.
Q2: How to query news from a specific date?
You can ask like this:
- "Query yesterday's news"
- "Check Zhihu news from 3 days ago"
- "What news was there on 2025-10-10"
- "News from last Monday"
Supported date formats: relative dates (today, yesterday, 3 days ago), days of week (last Monday, this Wednesday), absolute dates (2025-10-10).
Automatically queries today when date not specified. If you want to see all, request "show all news, don't summarize".
Q3: How to view trending topic statistics?
You can ask like this:
- "How many times did my followed words appear today" (using preset keywords)
- "Automatically analyze what hot topics are in today's news" (auto extract)
- "See what are the hottest words in the news" (auto extract)
Two modes: Preset keywords (count preset followed words based on config file, default) and Auto extract (auto-extract high-frequency words from news titles, no preset needed).
RSS Feed Queries (Q4.1-Q4.3)
Q4.1: How to view latest RSS feed content?
You can ask like this:
- "Show me the latest RSS feed content"
- "Get the latest articles from Hacker News"
- "View latest 20 items from all RSS feeds"
- "Get RSS feeds, need to include summaries"
- "Show me RSS content from the last week" (multi-day query support)
Returns today's RSS items by default (up to 50), supports
days parameter for multi-day queries (1-30 days). Does not include summaries by default, sorted by publication time in descending order, auto-deduplication across dates.Adjustable Specify RSS feed, specify days, adjust quantity, include summaries.
Q4.2: How to search content in RSS feeds?
You can ask like this:
- "Search for 'AI' related articles in RSS"
- "Search RSS content about 'machine learning' from last 7 days"
- "Search 'Python' in Hacker News"
Default searches last 7 days of data, returns up to 50 results. Adjustable Specify RSS feed, adjust days, include summaries.
Q4.3: How to view RSS feed status?
You can ask like this:
- "View RSS feed status"
- "How much data has RSS crawled"
- "Which RSS feeds have data"
Return information includes: available date list, total date count, today's per-feed statistics, generation time.
Search and Retrieval (Q4-Q5)
Q4: How to search for news containing specific keywords?
You can ask like this:
- "Search for news containing 'artificial intelligence'"
- "Find reports about 'Tesla price cut'"
- "Search for news about Musk, return top 20"
- "Find news about 'iPhone 16' in the last 7 days"
- "Find news about 'Tesla' from January 1 to 7, 2025"
- "Find the link to the news 'iPhone 16 release'"
Uses keyword mode search, default searches today's data, returns up to 50 results, does not include URL links by default. AI automatically converts relative time like "last 7 days", "last week" to specific date ranges.
Adjustable Specify time range (relative/absolute dates), specify platform, adjust sorting, include links.
Q4.4: How to search both hot news and RSS content simultaneously?
You can ask like this:
- "Search for 'AI' content, including RSS"
- "Find news about 'artificial intelligence', also search RSS subscriptions"
- "Search for 'Tesla', both hot news and RSS"
Hot news results and RSS results are displayed separately. Hot news sorted by rank/relevance, RSS sorted by publish time. Default returns 50 hot news + 20 RSS items.
Adjustable RSS count, search only hot news (default) or only RSS.
Q5: How to find related news?
You can ask like this:
- "Find news similar to 'Tesla price cut'" (today)
- "Find news related to 'AI breakthrough' from yesterday" (history)
- "Search for historical reports about 'Tesla' from last week" (history)
- "See if there are reports similar to this news in the last 7 days" (history)
Only queries today's data by default when time not specified. Supports preset values (yesterday, last week, last month) and custom date ranges. Similarity threshold 0.5 (adjustable), returns up to 50 results, sorted by similarity.
Adjustable Specify time, adjust threshold, include links.
Trend Analysis (Q6-Q7)
Q6: How to analyze topic heat trends?
You can ask like this:
- "Analyze the heat trend of 'artificial intelligence' in the last week"
- "See if 'Tesla' topic is a flash in the pan or sustained hot topic"
- "Detect which topics suddenly went viral today"
- "Predict potential hot topics coming up"
- "Analyze the lifecycle of 'Bitcoin' in December 2024"
Four analysis modes:
- Heat trend -- Track topic heat changes
- Lifecycle -- Complete cycle from emergence to disappearance
- Anomaly detection -- Identify suddenly viral topics
- Prediction -- Predict future hot topics
Default analyzes last 7 days of data, statistics by day granularity.
Q7: How to compare different platforms' attention to topics?
You can ask like this:
- "Compare different platforms' attention to 'artificial intelligence' topic"
- "See which platform updates most frequently"
- "Analyze which keywords often appear together"
Three insight modes:
- Platform Compare -- Compare platform attention (default mode)
- Activity Stats -- Count platform posting frequency
- Keyword Co-occurrence -- Analyze keyword associations (minimum frequency 3 times)
Sentiment Analysis & Aggregation (Q8-Q11)
Q8: How to analyze news sentiment?
You can ask like this:
- "Analyze today's news sentiment"
- "See if 'Tesla' related news is positive or negative"
- "Analyze different platforms' sentiment towards 'artificial intelligence'"
- "See the sentiment of 'Bitcoin' within a week, choose the top 20 most important"
Default analyzes today's data, returns up to 50 news items, sorted by weight. This tool returns AI prompts -- AI generates sentiment analysis reports based on them.
Adjustable Specify topic, specify time, adjust quantity.
Q9: How to get deduplicated cross-platform news?
You can ask like this:
- "Help me aggregate today's news, remove duplicates"
- "See which news is reported on multiple platforms"
- "Show me deduplicated hotspot news"
- "Which news are cross-platform hot topics"
Automatically identifies the same event reported by different platforms, merges similar news. Return information includes: representative title, covered platforms, platform count, cross-platform status, best rank, comprehensive weight.
Adjustable Specify time, adjust similarity threshold, specify platforms.
Q10: How to generate daily or weekly hotspot summaries?
You can ask like this:
- "Generate today's news summary report"
- "Give me a weekly hotspot summary"
- "Generate news analysis report for the past 7 days"
Supports daily summary and weekly summary.
Q11: How to compare hotspot changes across different periods?
You can ask like this:
- "Compare this week and last week's hotspot changes"
- "See what's different between this month and last month"
- "Analyze 'artificial intelligence' heat difference in two periods"
- "Compare platform activity changes"
Three comparison modes:
- Overview -- News count change, keyword change, TOP news comparison
- Topic shift -- Rising topics, falling topics, newly appeared topics
- Platform activity -- News count change by platform
Time period presets: Today/Yesterday, This week/Last week, This month/Last month, or custom date range.
System Management (Q12-Q14)
Q12: How to view system configuration?
You can ask like this:
- "View current system configuration"
- "Display configuration file content"
- "What platforms are available?"
- "What's the current weight configuration?"
Can query: available platform list, crawler configuration, weight configuration, notification configuration.
Q13: How to check system running status?
You can ask like this:
- "Check system status"
- "Is the system running normally?"
- "When was the last crawl?"
- "How many days of historical data?"
Return information: system version and status, last crawl time, historical data days, health check results.
Q13.1: How to check for version updates?
You can ask like this:
- "Check for version updates"
- "Is there a new version?"
- "Is the current version up to date?"
Will check both TrendRadar and MCP Server versions simultaneously, telling you the current installed version, latest available version, whether an update is needed, and update recommendations.
If GitHub access is slow, say "check version updates, use proxy http://127.0.0.1:10801".
Q14: How to manually trigger a crawl task?
You can ask like this:
- "Please crawl current Toutiao news" (temporary query)
- "Help me fetch latest news from Zhihu and Weibo and save" (persistent)
- "Trigger a crawl and save data" (persistent)
- "Get real-time data from 36Kr but don't save" (temporary query)
Two modes:
- Temporary Crawl (default) -- Only return data without saving
- Persistent Crawl -- Save to output folder
Default crawls all platforms, does not include URL links. Adjustable Specify platform, save data, include links.
Storage Sync (Q15-Q18)
Q15: How to sync data from remote storage to local?
You can ask like this:
- "Sync last 7 days data from remote"
- "Pull data from remote storage to local"
- "Sync last 30 days of news data"
Use cases: crawler deployed in the cloud, data stored remotely, MCP Server deployed locally needs to pull data from remote for analysis.
Return information: number of successfully synced files, skipped dates, failed dates and error info.
Prerequisites: need to configure remote storage in config file or set environment variables.
Q16: How to view storage status?
You can ask like this:
- "View current storage status"
- "What's the storage configuration"
- "How much data is stored locally"
- "Is remote storage configured"
Return information includes: local storage, remote storage, and pull configuration details.
Q17: How to view available data dates?
You can ask like this:
- "What dates are available locally"
- "What dates are in remote storage"
- "Compare local and remote data dates"
- "Which dates only exist remotely"
Three query modes:
- Local -- View local data dates only
- Remote -- View remote data dates only
- Compare (default) -- Compare both
Q18: How to parse natural language date expressions?
You can ask like this:
- "Parse what days 'this week' is"
- "What date range does 'last 7 days' correspond to"
- "Last month's date range"
- "Help me convert 'last 30 days' to specific dates"
This tool uses server-side precise time calculations to ensure all AI models get consistent date ranges. Supports Chinese and English.
Article Reading (Q19-Q20)
Q19: How to read the full content of a news article?
You can ask like this:
- "Help me read the content of this news: https://example.com/news/123"
- "Get the article body from this link"
- "Read the detailed content of this report"
Converts web pages to clean Markdown format via Jina AI Reader, automatically removes ads and noise.
Typical workflow:
- First use
search_news(include_url=True)to search news and get links - Then use
read_article(url=link)to read the article body - AI analyzes, summarizes, translates the Markdown content
Notes: 5-second interval between requests, Jina AI Reader free service (100 RPM limit), some paywalled pages may not be fully accessible. Default timeout 30 seconds, max 60 seconds.
Q20: How to batch read multiple articles?
You can ask like this:
- "Help me read the content of these news articles"
- "Batch get the article bodies from these links"
- "Read the detailed content of the first 3 search results"
Typical workflow:
- First use
search_news(include_url=True)to search news and get multiple links - Then use
read_articles_batch(urls=[...])to batch read article bodies - AI performs comparative analysis, comprehensive reports on multiple articles
Tool limits: max 5 articles per batch, 5-second request interval, estimated time (5 articles) 25-30 seconds.
Notification Push (Q21)
Q21: How to send notification messages via MCP?
You can ask like this:
- "Show me which notification channels are configured"
- "Send a test message to all channels"
- "Push this content to Feishu"
- "Send today's news summary to DingTalk and Telegram"
Supports 9 notification channels: Feishu, DingTalk, WeCom, Telegram, Email, ntfy, Bark, Slack, Generic Webhook.
Configuration: configure channels in
config.yaml under notification.channels, or set environment variables in .env file (higher priority). Both sources are automatically merged.Typical workflow:
- Check channel status first
- Send after confirming availability
- Or specify multiple channels
- Without specifying channels, sends to all configured channels
The tool accepts Markdown format messages and automatically converts to each channel's required format. Multi-account support: separate multiple URLs/Tokens with
; in config values.Usage Tips
How to make AI display all data instead of auto-summarizing?
Sometimes AI automatically summarizes data, only showing partial content. If AI still summarizes, you can:
- Method 1 - Explicit request: "Please show all news, don't summarize"
- Method 2 - Specify quantity: "Show all 50 news items"
- Method 3 - Question the behavior: "Why only showed 15? I want to see all"
- Method 4 - State upfront: "Query today's news, fully display all results"
Note: AI may still adjust display method based on context.
How to combine multiple tools?
Example: In-depth analysis of a topic
- Search first: "Search for news about 'artificial intelligence'"
- Then analyze trends: "Analyze the heat trend of 'artificial intelligence'"
- Finally sentiment analysis: "Analyze sentiment of 'artificial intelligence' news"
Example: Track an event
- View latest: "Query today's news about 'iPhone'"
- Find history: "Find historical news related to 'iPhone' from last week"
- Find similar reports: "Find news similar to 'iPhone launch event'"