@Honcia13: Never worry about missing hot topics scrolling through dozens of apps again! This 50k-star open-source tool TrendRadar directly aggregates hot topics from across the web, uses AI to filter precisely based on your interests, pushing only relevant content to your phone! Supports keywords + natural language descriptions like “what I want to see”, AI auto-translation + briefing generation, one-click push to…

Never have to refresh dozens of apps anxiously fearing you’ll miss hot topics again! This open-source gem TrendRadar with 50k stars aggregates hot topics from across the web, uses AI to filter precisely based on your interests, and pushes only content relevant to you straight to your phone! Supports keywords + natural language descriptions of “what I want to see”, AI automatic translation + briefing generation, one-click push to WeChat/Feishu/Telegram and 9 major channels. 30-second Docker deployment, complete local data sovereignty! Highly recommended by Yifeng Ruan and XiaoZhong Software! A direct salvation for those suffering from information anxiety https://github.com/sansan0/TrendRadar… — # sansan0/TrendRadar Source: https://github.com/sansan0/TrendRadar Hotspot assistant deployed in as little as 30 seconds —— Say goodbye to无效刷屏，only read news and information you truly care about GitHub Stars GitHub Forks License Version (https://github.com/sansan0/TrendRadar) MCP (https://github.com/sansan0/TrendRadar) RSS (https://github.com/sansan0/TrendRadar) AI Translation (https://github.com/sansan0/TrendRadar) WeCom Notification (https://work.weixin.qq.com/) Personal WeChat Notification (https://weixin.qq.com/) Telegram Notification (https://telegram.org/) DingTalk Notification Feishu Notification (https://www.feishu.cn/) Email Notification ntfy Notification (https://github.com/binwiederhier/ntfy) Bark Notification (https://github.com/Finb/Bark) Slack Notification (https://slack.com/) Generic Webhook GitHub Actions (https://github.com/sansan0/TrendRadar) GitHub Pages (https://sansan0.github.io/TrendRadar) Docker (https://hub.docker.com/r/wantcat/trendradar) MCP Support (https://modelcontextprotocol.io/) AI Analysis Push AI Smart Filtering Chinese | English > This project aims for lightweight design and easy deployment ## 📑 Quick Navigation > 💡 Click the links below to quickly jump to corresponding sections. Deployment is recommended to start from “Quick Start**, and for detailed customization please refer to “Configuration Details” | | | | |:—:|:—:|:—:| | 🚀 Quick Start | AI Smart Analysis | ⚙️ Configuration Details | | Docker Deployment | MCP Client | 📝 Change Log | | 🎯 Core Features | ☕ Support Project | 📚 Project Related | - Thank you to all viewers who starred the project. Forking is what you want, starring is what I want, having both 😍 is the best support for the open-source spirit 👉 Click to expand: List of Acknowledgements (Seed Round Honor Roll 🔥73+🔥 people) ### Acknowledgement of Early Supporters > 💡 Special Note: > > 1. About the List: The table below records supporters during the startup phase (Seed Round) of the project. Due to the tedious nature of manual statistics early on, there may inevitably be omissions or incomplete records; if anyone is missed, it was unintentional, please understand. > 2. Future Plan: To focus limited energy on code and feature iteration, manual maintenance of this list will cease effective immediately. > > Regardless of whether a name is on the list, every bit of your support is the cornerstone that allowed TrendRadar to reach where it is today. 🙏 ### Infrastructure Support Thanks to GitHub for providing free infrastructure, which is the biggest premise allowing this project to run conveniently with one-click fork. ### Data Support This project uses the API of the newsnow (https://github.com/ourongxing/newsnow) project to obtain multi-platform data; special thanks to the author for providing the service. After contact, the author stated there is no need to worry about server pressure, but this is based on his kindness and trust. Please everyone: - Go to the newsnow project (https://github.com/ourongxing/newsnow) and star it to show support - When deploying via Docker, please reasonably control the push frequency, do not drain the pond to catch fish (avoid overusing resources). ### Promotion Support > Thanks to the following platforms and individuals for recommendations (listed by time) - XiaoZhong Software (https://mp.weixin.qq.com/s/fvutkJ_NPUelSW9OGK39aA) - Open Source Software Recommendation Platform - LinuxDo Community (https://linux.do/) - Gathering place for tech enthusiasts - Yifeng Ruan Weekly (https://github.com/ruanyf/weekly) - Influential weekly newsletter in the tech circle ### Audience Support > Thanks to friends who provided financial support. Your generosity has transformed into snacks and drinks next to the keyboard, accompanying every iteration of the project. > > Regarding the return of “One-Yuan Appreciation”: > With the release of version v5.0.0, the project has entered a new stage. To support the growing API costs and caffeine consumption, the “One-Yuan Appreciation” channel has reopened. Every bit of your intention will be converted into Tokens and momentum in the code world. 🚀 Go to Support | Praisier | Amount | Date | Remark | | :———————––: | :––: | :––: | :———————–: | | D*5 | 1.8 * 3 | 2025.11.24 | | | Ghost | 1 | 2025.11.17 | | | Super | 10 | 2025.11.17 | | | Rw | 10 | 2025.11.17 | This agent is amazing, brother | | Jo | 1 | 2025.11.17 | Thanks for open source, wish you success | | *Morning | 8.88 | 2025.11.16 | Good project, researching and learning | | *Sea | 1 | 2025.11.15 | | | De | 1.99 | 2025.11.15 | | | Shu | 8.8 | 2025.11.14 | Thanks for open source, great project, supporting | | Me | 10 | 2025.11.14 | Open source is hard, boss you worked hard | | **Ke | 1 | 2025.11.14 | | | Cloud | 88 | 2025.11.13 | Great project, thanks for open source | | W | 6 | 2025.11.13 | | | Kai | 1 | 2025.11.13 | | | To. | 1 | 2025.11.13 | Thanks for your TrendRadar | | sy | 1 | 2025.11.13 | | | **Xiang | 10 | 2025.11.13 | Great project, found you too late, thanks for open source! | | Wei | 9.9 | 2025.11.13 | TrendRadar is awesome, buy teacher a coffee~ | | hp | 5 | 2025.11.12 | Supporting Chinese open source power, come on! | | cr | 6 | 2025.11.12 | | | an | 5 | 2025.11.12 | | | . *c | 1 | 2025.11.12 | Thanks for sharing open source | | *Ji | 1 | 2025.11.11 | | | *Master | 1 | 2025.11.10 | | | Le | 10 | 2025.11.09 | | | Jie | 5 | 2025.11.08 | | | Dian | 8.80 | 2025.11.07 | Development is hard, supporting. | | QQ | 6.66 | 2025.11.07 | Thanks for open source! | | Ce | 1 | 2025.11.05 | | | Peter Fan | 20 | 2025.10.29 | | | Mn | 1 | 2025.10.27 | Thanks for open source | | Xu | 8.88 | 2025.10.23 | Teacher, I’m a newbie, touched it for days still haven’t set it up, seeking advice | | Eason | 1 | 2025.10.22 | Haven’t figured it out yet, but you are doing good things | | Pn | 1 | 2025.10.20 | | | *Jie | 1 | 2025.10.19 | | | *Xu | 1 | 2025.10.18 | | | *Zhi | 1 | 2025.10.17 | | | *😀 | 10 | 2025.10.16 | Thumbs up | | **Jie | 10 | 2025.10.16 | | | *Xiao | 10 | 2025.10.16 | | | Ji | 5 | 2025.10.14 | TrendRadar | | Jd | 1 | 2025.10.14 | Thanks for your tool, very fun… | | H | 1 | 2025.10.14 | | | NaO | 10 | 2025.10.13 | | | Yuan | 1 | 2025.10.13 | | | Pg | 6 | 2025.10.13 | | | Ocean | 20 | 2025.10.12 | …Really too amazing!!! Even beginners can use it directly… | | **Pei | 5.2 | 2025.10.2 | github-yzyf1312: Long live open source | | *Chun | 3 | 2025.9.23 | Come on, very nice | | 🍍 | 10 | 2025.9.21 | | | Ef | 1 | 2025.9.20 | | | Ji | 1 | 2025.9.20 | | | zu | 2 | 2025.9.19 | | | **Hao | 5 | 2025.9.17 | | | Hao | 1 | 2025.9.15 | | | TT | 2 | 2025.9.15 | Thumbs up | | *Jia | 10 | 2025.9.10 | | | X | 1.11 | 2025.9.3 | | | Biao | 20 | 2025.8.31 | From Old Tong thanks | | Xia | 1 | 2025.8.30 | | | 2D | 88 | 2025.8.13 PM | | | 2D | 1 | 2025.8.13 AM | | | So | 1 | 2025.8.05 | Supporting | | Knight | 10 | 2025.8.04 | | | xx | 2 | 2025.8.03 | trendRadar great project thumbs up | | *Yuan | 1 | 2025.8.01 | | | *Xie | 5 | 2025.8.01 | | | *Meng | 0.1 | 2025.7.30 | | | **Dragon | 10 | 2025.7.29 | Supporting | ## 🪄 Sponsors > Vacancy Available ### ❤️ Think it’s useful? Support us > If TrendRadar has ever captured value for you, consider injecting momentum into it to help it continue evolving > > Amount is optional, even 1 Yuan is encouragement for open source. Welcome to leave a remark when appreciating ( ́▽ʃ♡ƪ) | WeChat Appreciation | Alipay Appreciation | |:---:|:---:| | | | ### 🤝 Secondary Development and Citation If you use or borrow ideas/core code from this project in your own work, it is **highly welcomed** to indicate the source and include a link to this repository in the README or documentation. This will help with the continued maintenance and community development of the project; thank you for your respect and support! ❤️ ### 💬 Communication and Feedback - **GitHub Issues**: Suitable for specific technical questions. When asking, please provide complete information (screenshots, error logs, etc.) to help locate issues quickly. - **Official Account Interaction**: Recommended to communicate first in the comment section under related articles. If you need to ask via the backend, **liking/recommending** the article first is the best "foot-in-the-door"; I can feel this thoughtfulness in the backend ( ́▽ʃ♡ƪ). - QQ Group Interaction: Follow the official account and reply with “Exchange Group” to join. Whether you are an AI beginner or a hardcore developer, seeking technical help or sharing experiences, you are welcome here. The group focuses on mutual assistance and inspiration collision; please read the group announcement upon joining; describe issues clearly and attach screenshots when asking, and group members will help when available; everyone’s practical experience is often faster and more comprehensive than mine alone 🤝 > Friendly Reminder: > This project is for open-source sharing, not a commercial product. Treating the author as a friend rather than customer service will result in higher communication efficiency! | Follow Official Account | |:—:| | | ## 📝 Change Log > 📌 View Latest Updates:Original Repo Change Log (https://github.com/sansan0/TrendRadar?tab=readme-ov-file#-更新日志) : - Hint: Suggested to view [Historical Updates] to clarify specific [Feature Content] ### 2026/03/28 - v6.6.0 - HTML Report Browser Enhancement: Opening reports in the browser automatically switches to wide-screen layout. Keyword grouping and independent display zones both support Tab quick switching. Search box filters news titles in real-time. Email clients still display the original narrow-screen layout, zero regression. - Dark Mode: One-click switch to dark theme, preferences remembered automatically, suitable for night reading. - One-Click Copy News: Hover mouse over news number to copy title and link, convenient for quick sharing. - Export Optimization: Full-page screenshot and segmented screenshot merged into a dropdown export button. Layout automatically restored cleanly during screenshot. - Shortcut Key System: Supports W for Wide-screen toggle, D for Dark Mode, / for Search, ? for viewing shortcut hints. - Reading Progress Bar: Reading progress displayed in real-time at top of page. ### 2026/02/09 - mcp-v4.0.0 - 🔥 AI Message Direct Push to All Channels: Let AI-written content be pushed to Feishu, DingTalk, Telegram, Email, and other 9 channels with one click. Markdown automatically adapts to each platform’s format; no need to worry about format differences. - New Formatting Strategy Guide: Added get_channel_format_guide tool, tells AI what formats each channel supports and what restrictions exist, making generated content layout look better. - Smart Batch Sending: Ultra-long messages automatically split according to byte limits of each channel (Feishu 30KB, DingTalk 20KB, etc.), configuration read from config.yaml. - Fixed Channel MisDetection: ntfy no longer reported as “configured” due to default address. - Code Reuse Optimization: Batch processing functions directly reuse trendradar core modules, no reinventing the wheel. 👉 Click to expand: Historical Updates ### 2026/03/12 - v6.5.0 - AI Intelligent Screening System: No need to manually set keywords anymore! Write your areas of interest in everyday language in ai_interests.txt (e.g., “I want to see AI and New Energy related news”), AI will automatically extract tags and score each news item, pushing only content truly relevant to you. If AI screening encounters issues, it automatically reverts to keyword matching; push does not interrupt. - Different Screening Methods and Focus Areas Supported Per Time Slot: Each time period in Timeline can now independently set screening method and news type to view. E.g.: Morning uses “Tech Keywords” for quick filtering, Evening switches to “Finance AI Interest Description” for deep screening — same system, different content viewed at different times. - AI Analysis Scope Independent from Push: AI analysis data scope can differ from push content. E.g.: Push only sends new messages (avoid repeat disturbance), but AI analyzes all day’s news (view complete trends). Each time slot can also individually set AI analysis mode. - AI Screening Cost Savings: Analyzed news will not consume tokens repeatedly; if interest description is modified, AI automatically judges change magnitude — small tweaks update only affected tags, large changes trigger full reclustering. - Multi-File Configuration and Tag Isolation: Custom keyword files placed in config/custom/keyword/, AI interest files in config/custom/ai/. Tags generated from different files are independent and non-interfering. - AI Translation Precise Control: Can separately control whether Hot Rankings, RSS, Independent Display Zone should be translated. Disabled display areas automatically skipped, saving tokens. - Remote Storage Batch Upload: Multiple write operations accumulated and submitted to cloud together, reducing API call count. - News Count Limit Per Keyword/Tag Group: Control max number of news displayed per group via max_news_per_keyword, avoiding single hot topic filling entire push. - Time Slot Conflict Smart Detection: If two time slots overlap, system automatically errors and reminds modification, avoiding configuration conflict leading to unexpected behavior. - Fixed several bugs ### 2026/02/09 - v6.0.0 > Breaking Change: Configuration file upgrade (config.yaml 2.0.0), old push_window and analysis_window configs no longer compatible, please refer to new config.yaml for migration. - Unified Scheduling System: Added timeline.yaml, control “when to collect/push/AI analyze” with one config set. - 5 Preset Templates: always_on(All-day, default), morning_evening(AM/PM Summary), office_hours(Office Hours), night_owl(Night Owl), custom(Custom); Also supports adding your own templates under presets: as long as key is unique, then fill template name in config.yaml. - Flexible Time Slot Configuration: Supports weekday/weekend differentiation, cross-midnight time slots, per-period once deduplication. - Visual Configuration Editor: - Added timeline.yaml edit tab, alongside config.yaml / frequency_words.txt - Preset mode card selection: Click to switch, auto syncs config.yaml’s schedule.preset - Week View Timeline: 7 days × 24 hours horizontal bar, uses colors to distinguish Push/Analysis/Collection status - Interactive Controls: Switches, dropdowns, time pickers, right-side edits sync real-time to left YAML - Week Mapping Dropdown Select: Dynamically fills based on daily plan, drag/click to complete scheduling config - AI Prompt Stability Optimization(ai_analysis_prompt.txt v2.0.0): - Format Spec Independent Instructions: Extracted line breaks/tags/numbers/prohibitions from JSON values into independent chapters - JSON Template Simplified: Field descriptions shortened to one sentence + char limit, reduces AI output format chaos - Removed Markdown formatting from system prompt, consistent with “No Markdown” instruction - All JSON fields declared optional, missing any field won’t error, enhances fault tolerance - Added Independent Display Zone AI Summary Analysis(ai_analysis.include_standalone): - Added standalone switch, enabling AI to generate core summary for each standalone source - AI Analysis Decoupled from Push Display: No need to enable push display for Independent Display Zone, AI can independently analyze complete Hot Ranking data - Supports Hot Ranking Platforms and RSS sources, includes ranking/time/trajectory data - Trajectory Analysis Linked with include_rank_timeline: When enabled utilizes trajectory data for deep trend analysis, when off based on ranking for brief judgment - Added standalone_summaries JSON field (Standalone Source Snapshot), all push channels adapted for rendering ### 2026/01/28 - v5.5.0 > Same as mcp functionality, I won’t create a new repo to maintain this small tool either, purely frontend, put it all together. - Added TrendRadar visual configuration editor ### 2026/02/02 - mcp-v3.2.0 - New read_article Tool: Reads single article body via Jina AI Reader (Markdown format). - New read_articles_batch Tool: Batch reads multiple articles (max 5, auto rate limiting). - Recommended Workflow: search_news(query="Keyword", include_url=True) → read_article(url=...) to read body. - Docs Update: README-MCP-FAQ.md and README-MCP-FAQ-EN.md added Q19-Q20 article reading instructions. ### 2026/01/10 - mcp-v3.0.0~v3.1.5 - Breaking Change: All tool return values unified to {success, summary, data, error} structure. - Async Consistency: All 21 tool functions wrapped with asyncio.to_thread() for synchronous calls. - MCP Resources: Added 4 resources (platforms, rss-feeds, available-dates, keywords). - RSS Enhancement: get_latest_rss supports multi-day query (days parameter), cross-date URL deduplication. - Regex Match Fix: get_trending_topics supports /pattern/ regex syntax and display_name. - Cache Optimization: Added make_cache_key() function, param sorting + MD5 hash ensures consistency. - New check_version Tool: Supports simultaneous check for TrendRadar and MCP Server version updates. ### 2026/01/23 - v5.4.0 - Added independent control feature for AI analysis modes, optional follow_report | daily | current | incremental. - Added AI analysis time window control, supports custom running segments and daily frequency limits. - Added configuration file version management functionality. - Fixed several bugs. ### 2026/01/19 - v5.3.0 > Major Refactor: AI Module Migrated to LiteLLM - Unified AI Interface: Using LiteLLM replacing manual implementation, supports 100+ AI providers. - Simplified Config: Removed provider field, switched to model: "provider/model_name" format. - New Functionality: Auto retry (num_retries), fallback models (fallback_models). - Config Changes: - ai.provider → Removed (merged into model) - ai.base_url → ai.api_base - AI_PROVIDER env var → Removed - AI_BASE_URL env var → AI_API_BASE - Model Format Examples: - DeepSeek: deepseek/deepseek-chat - OpenAI: openai/gpt-4o - Gemini: gemini/gemini-2.5-flash - Anthropic: anthropic/claude-3-5-sonnet ### 2026/01/17 - v5.2.0 > Mainly see config.yaml description 🌐 AI Translation Feature - Multi-language Translation: Supports translating push content to any language. - Batch Translation: Smart batch processing, reduces API call count. - Custom Prompts: Supports custom translation styles. 🔧 Config Architecture Optimization - AI Model Config Independent: Analysis and Translation share model config. - Region Switch Unified: Unified management of push region display. - Region Sort Custom: Supports customizing display order of regions. ✨ AI Analysis Enhancement - AI Analysis Embedded HTML: Analysis results directly embedded in HTML report, email notification uses directly. - Rich Style AI Block: Gradient blue background card layout, clearly separates each analysis dimension. - Ranking Timeline Support: AI can get exact ranking for each news item at every crawl timestamp. - Sector Restructure (7→4): Integrated into Core Hotspot态势, Public Opinion Controversy, Anomalies & Weak Signals, Judgment Strategy Suggestions. 🔧 Multi-Model Adaptation - Universal Parameter Pass-through: Supports passing arbitrary advanced parameters to API. - Gemini Adaptation: Native parameter support, built-in security policy relaxation. 🐛 Bug Fixes - Fixed several known issues, improved system stability. ### 2026/01/10 - v5.0.0 > Development Interlude: > Paying tribute to a certain major model provider that accompanied me for over two years, but popped up "This organization has been disabled" right after renewal. ✨ Push Content “Five Major Sections” Restructure This update restructured push messages regionally. Now push content is clearly divided into five core sections: 1. 📊 Hot Ranking News: Web-wide hotspot aggregation precisely filtered by your keywords. 2. 📰 RSS Subscription: Your personalized subscription source content, supports grouping by keywords. 3. 🆕 Newly Added: Real-time capture of entirely new hotspots since last run (marked with 🆕). 4. 📋 Independent Display Zone: Complete hot rankings or RSS sources from specified platforms, completely unrestricted by keyword filtering. 5. ✨ AI Analysis Section: Deep insights driven by AI, including trend overview, heat trajectory, and extremely important sentiment tendency analysis. ✨ AI Intelligent Analysis Push Feature - AI Analysis Integration: Uses AI Large Models to deeply analyze push content, automatically generates hotspot trend overviews, keyword heat analysis, cross-platform associations, potential impact assessments, etc. - Sentiment Tendency Analysis: Added deep sentiment recognition, accurately capturing public opinion positive/negative, controversy, or concern emotions. - Multi-AI Provider Support: Supports DeepSeek (default, high cost-performance ratio), OpenAI, Google Gemini, and any OpenAI-compatible interface. - Two Push Modes: only_analysis (AI Analysis only), both (Both push). - Custom Prompts: Customize AI analysis role and output format via config/ai_analysis_prompt.txt file. - Multi-dimensional Data Analysis: AI can analyze ranking changes, heat duration, cross-platform performance, trend predictions, etc. 📋 Independent Display Zone Feature - Complete Hot Ranking Display: Complete hot rank of specified platforms displayed separately, unaffected by keyword filtering. - RSS Independent Display: RSS source content can be displayed completely, suitable for subscription sources with less content. - Flexible Configuration: Supports configuring display platform list, RSS source list, max display count. 📊 Push Experience Restructure - Layout Upgrade: Redesigned and unified stats header for each channel, strengthened block organization, message hierarchy clear at a glance. - Config Simplification: Optimized configuration logic for Feishu and other notification channels, easier to get started. - Heat Trend Arrows: Added 🔺 (Rising), 🔻 (Falling), ➖ (Flat) trend indicators, intuitively displaying heat changes. - Generic Webhook: Supports custom Webhook URL and JSON template, easily adapts to Discord, Matrix, IFTTT, or any platform. 🔧 Config Optimization - Frequency Word Config Enhancement: Added [Group Alias] syntax, supports # comment lines, config clearer (Thanks to suggestion from @songge8 (https://github.com/sansan0/TrendRadar/issues/752)). - Env Variable Support: AI analysis related configs support env variable override (AI_API_KEY, AI_PROVIDER, etc.). > 💡 Detailed config tutorial see Let AI Help Me Analyze Hotspots ### 2026/01/02 - v4.7.0 - Fixed RSS HTML Display: Fixed rendering issues caused by mismatched RSS data formats, now groups by keyword display correctly. - New Regex Syntax: Keyword config supports /pattern/ regex syntax, solves false match issues with English substrings (e.g. ai matches training) 📖 View Syntax Details. - New Display Name Syntax: Use => Remark to give complex regex expressions memorable names, push messages display clearer (e.g. /\bai\b/ => AI Related). - Don’t know regex? README added guidance for AI-generated regex, tell ChatGPT/Gemini/DeepSeek what you want to match, let AI help write it. ### 2025/12/30 - mcp-v2.0.0 - Architecture Adjustment: Removed TXT support, unified to SQLite database. - RSS Query: Added get_latest_rss, search_rss, get_rss_feeds_status. - Unified Search: search_news supports include_rss parameter to search Hot Rankings and RSS simultaneously. ### 2026/01/01 - v4.6.0 - Fixed RSS HTML Display: Merged RSS content into Hot Ranking HTML pages, grouped by source display. - New display_mode Config: Supports keyword (group by keyword) and platform (group by platform) display modes. ### 2025/12/30 - v4.5.0 - RSS Subscription Support: Added RSS/Atom crawling, grouped by keyword stats (consistent with Hot Ranking format). - Storage Structure Refactor: Flattened directory structure output/{type}/{date}.db. - Unified Sort Config: sort_by_position_first affects both Hot Rankings and RSS. - Config Structure Refactor: config.yaml reorganized into 7 logical groups (app, report, notification, storage, platforms, rss, advanced), config paths clearer. ### 2025/12/26 - mcp-v1.2.0 MCP Module Update - Optimized Toolset, Added Aggregation Comparison, Merged Redundant Tools: - Added aggregate_news tool - Cross-platform news deduplication aggregation. - Added compare_periods tool - Period comparison analysis (Week-over-week / Month-over-month). - Merged find_similar_news + search_related_news_history → find_related_news. - Enhanced get_trending_topics - Added auto_extract mode for automatic hotspot extraction. - Fixed several bugs. - Sync updated README-MCP-FAQ.md document Chinese and English versions (Q1-Q18). ### 2025/12/20 - v4.0.3 - Added URL normalization feature, solves repeated push issues caused by dynamic params (e.g. band_rank) on platforms like Weibo. - Fixed incremental mode detection logic, correctly identifies historical titles. ### 2025/12/17 - v4.0.1 - StorageManager added push record proxy methods. - S3 Client switched to virtual-hosted style to improve compatibility (supports Tencent Cloud COS and more services). ### 2025/12/13 - mcp-v1.1.0 MCP Module Update: - Adapted for v4.0.0, also compatible with v3.x data. - Added storage sync tools: sync_from_remote, get_storage_status, list_available_dates. ### 2025/12/13 - v4.0.0 🎉 Major Update: Comprehensive Refactor of Storage and Core Architecture - Multi-Storage Backend Support: Introduced new storage module, supports local SQLite and remote cloud storage (S3 compatible protocol, e.g. Cloudflare R2), adapts to GitHub Actions, Docker, and local environments. - Database Structure Optimization: Refactored SQLite database table structure, improved data efficiency and query capability. - Core Code Modularization: Split main program logic into multiple modules of trendradar package, significantly improved code maintainability. - Enhanced Features: Implemented date format standardization, data retention policies, timezone config support, time display optimization, fixed remote storage data persistence issues, ensured accuracy of data merging. - Cleanup and Compatibility: Removed most historical compatibility code, unified data storage and retrieval methods. ### 2025/12/03 - v3.5.0 🎉 Core Feature Enhancement 1. Multi-Account Push Support - All push channels (Feishu, DingTalk, WeCom, Telegram, ntfy, Bark, Slack) support multi-account config. - Use semicolon ; to separate multiple accounts, e.g.: FEISHU_WEBHOOK_URL=url1;url2. - Automatic validation of paired configs (e.g. Telegram token and chat_id) quantity consistency. 2. Push Region Config - Customize display order of regions via display.region_order (v5.2.0 replaces original reverse_content_order). - Control whether each region displays via display.regions (Hot Rankings, New Hotspots, RSS, Independent Display, AI Analysis). 3. Global Filter Keywords - Added [GLOBAL_FILTER] area marker, supports global filtering of unwanted content. - Applicable scenarios: Filtering ads, marketing, low-quality content, etc. 🐳 Docker Dual-Path HTML Generation Optimization - Issue Fix: Resolved issue where index.html failed to sync to host in Docker environment. - Dual-Path Generation: Daily summary HTML generated to two locations simultaneously - index.html (Project root dir): For GitHub Pages access - output/index.html: Mounted via Docker Volume, host accessible directly. - Compatibility: Ensures Docker, GitHub Actions, Local Run Environment can all access web report normally. 🐳 Docker MCP Image Support - Added standalone MCP service image wantcat/trendradar-mcp. - Supports Docker deployment of AI analysis features, serving via HTTP interface (port 3333). - Dual Container Architecture: News push service and MCP service run independently, can scale/restart separately. - See Docker Deployment - MCP Service. 🌐 Web Server Support - Added built-in Web server, supports accessing generated reports via browser. - Controlled via manage.py command start/stop: docker exec -it trendradar python manage.py start_webserver. - Access Address: http://localhost:8080 (port configurable). - Security Features: Static file service, directory restriction, local access. - Supports auto-start and manual control modes. 📖 Documentation Optimization - Added How Push Content Displays? section: Customize push style and content. - Added When Do You Push to Me? section: Set push time slots. - Added How Often Run? section: Set auto-run frequency. - Added Push to Multiple Groups/Devices section: Push to multiple recipients simultaneously. - Optimized config sections: Uniformly added “Config Location” notes. - Simplified Quick Start config notes: Three core files clear at a glance. - Optimized Docker Deployment section: Added image notes, recommended git clone deployment, reorganized deployment methods. 🔧 Upgrade Notes: - GitHub Fork Users: Update main.py, config/config.yaml (added multi-account push support, no need to modify existing config). - Multi-Account Push: New feature, not enabled by default, existing single-account config unaffected. ### 2025/11/26 - mcp-v1.0.3 MCP Module Update: - Added date parsing tool resolve_date_range, solved inconsistent date calculation by AI models. - Supports natural language date expression parsing (this week, last 7 days, last month, etc.). - Total tool count increased from 13 to 14. ### 2025/11/28 - v3.4.1 **