sansan0
[建议] 德国？友人希望本项目支持 whatsapp 推送？
🏷️ 建议类别
数据源相关（新增平台、改进抓取等）
💭 详细描述你的想法
我考虑添加WhatsApp以及一套基于欧洲（德国）和英语的资源。我猜我的第一步会是新增通知路径。我只会说50个中文单词（我家孩子说得流利）。中期我能用微信。我可能还会添加一些英文设置文档。问候。
🎯 什么时候会用到这个功能
实施之后
🛠️ 实现想法（可选）
增加更多通知渠道，添加爬虫和新闻来源，产出更多文档
📷 功能示意图（推荐）
No response
📎 其他补充说明
No response
i did some stuff in my fork including a test suite, timezone support and stuff. well copilot did most of the stuff ;)
i removed the cached output but just because i renamed the folder structure. have a look : https://github.com/qvest-ssels/TrendRadar/tree/develop
the most important thing for this repo's i18n is more international news source, i think. however, the news pulling in this repo depends on newsnow, which is still a demo only supporting Chinese. that's so bad (sigh) i believe many China user also wanna more international news source like me.
isn't there any other repo like newsnow supporting it? i don't believe newsnow is the only one to crawl news. Maybe because it's hard to bypass the paywall for international newspaper website such as wsj and so on?
there are paywall cache services: https://archive.is/ but those well. are 复印 sources ;) i gonna have another look about alternative crawling services. i added some simple RSS support to my develop branch
my fork has some more nice features i am implementing including an chatbot, translation services and static pages with headlines. all done by copilot btw with some nice prompting
What's difference between the chatbot and AI agent with origin repo's MCP? I can use Cherry Studio to do the same thing.
the crawler also is able to crawl cloudflare protected sites for headlines and use playwright to parse javascript and firefox (with an external lib etc) readermode to filter/transform news articles.
That's nice!
Share my screenshot using cherry studio with grok 4.1 fast (free model from openrouter)
that looks also nice. how do you do the translation? i have also an local translater if needed included. via https://github.com/LibreTranslate/LibreTranslate
i also extended the MCP to be able to do searches on the website not only RSS and static pages. this feature is pretty fresh though
I don't know. There is a button and I clicked it. (lol) It seems not related to LLM because it doesn't show how many tokens used for translating.
feel free to clone the fork and play around the python code is purely gernerated by github copilot using claude opus. i did not write any line besides prompts and instructions.
for the local chat i use an M2 24GB notebook, not tryed intel/amd64 architecture yet.