sansan0/TrendRadar

TrendRadar – AI‑enhanced news‑hotspot assistant

What it is

• An open‑source tool that collects the latest news and hot‑list items from many platforms (via the newsnow API) and pushes a concise, filtered digest to you.
• It adds optional AI‑driven features: automatic translation, smart keyword‑based filtering, and deep analysis of trends, sentiment and cross‑platform signals.

Why it matters

• Keeps you up‑to‑date without the noise of endless scrolling – you only see the stories you actually care about.
• The AI layer can turn raw headlines into short summaries, translate them into any language, and even generate a brief “trend report” that highlights rising topics, sentiment shifts and potential impact.

Key features (as described in the README)

Feature	What you get
Fast deployment	One‑click Docker image (wantcat/trendradar) or local Python install – the README claims a 30‑second setup.
Multi‑channel notifications	Sends the digest to Feishu, DingTalk, Telegram, Email, Slack, WeChat (work & personal), ntfy, Bark, generic Webhook, etc.
AI translation	Batch‑processes headline translation to avoid rate‑limit errors; supports any target language.
AI smart filtering (ai_interests.txt)	Write natural‑language interests (e.g. “AI and renewable‑energy news”). The AI extracts tags, scores each article and only pushes relevant items.
AI analysis	Generates a concise “trend overview” with sentiment, cross‑platform correlation, and suggestions. The analysis can be sent separately or together with the news list.
Configurable schedule (timeline.yaml)	Define when to fetch, push and run AI analysis – presets for always‑on, morning/evening, office hours, night‑owl, or custom schedules.
HTML report	A browsable web page with dark‑mode, tabbed sections, trend arrows, copy‑click headlines, and export to Markdown or screenshots.
RSS support	Pulls user‑defined RSS feeds, de‑duplicates via GUID, and can display them as a separate block.
Self‑hosted data source	You can run your own newsnow instance and point TrendRadar at it.
MCP client	A set of async Python tools (21 functions) for programmatic access – e.g. search_news, read_article, version checks, etc.
Extensible	Plug‑in‑style configuration, custom prompts for AI, and a visual config editor in the web UI.

How to get started

1. Deploy – pull the Docker image (docker pull wantcat/trendradar) or run the Python package locally.
2. Configure – edit config.yaml (or use the visual editor) to set your notification channels, keywords, AI model (OpenAI, Gemini, DeepSeek, etc.) and schedule.
3. Run – the service will fetch hot‑list data, apply AI filtering/translation, generate the HTML report and push the digest according to your schedule.

Who might use it

• Professionals who need a daily briefing of industry‑specific news.
• Developers who want a ready‑made pipeline for collecting, translating and analysing trending content.
• Communities that want to broadcast curated news via chat tools (WeChat, Telegram, Slack, etc.).

License & community

• Licensed under GPL‑3.0.
• Actively maintained (v6.10.0 as of June 2026) with frequent releases, Docker pulls, and a public issue tracker.
• The project acknowledges sponsors and offers a “SanFrame” companion tool for clipboard and OCR management.

---

All details above are taken directly from the repository’s README; no additional assumptions have been made.