[Feature Request] Add arXiv paper retrieval by title keywords for daily papers & content summarization 【功能建议】新增 arXiv 按配置标题关键词获取当日论文及内容总结功能

[meal-maker]

1. Is your feature request related to a problem? 该功能需求是否与现有问题相关？

Currently the project does not support obtaining the latest papers from arXiv by pre-set title keywords, nor can it automatically summarize paper content. It is not convenient for users to track and quickly grasp key information of arXiv papers published on the day.

目前项目暂不支持通过预设标题关键词从 arXiv 获取当日最新论文，也无法对论文内容进行自动概括总结，不利于用户追踪并快速掌握当日 arXiv 相关论文的核心信息。

2. Describe the solution you'd like. 你希望实现的解决方案

Add a new feature based on the official arXiv API:

• Support configuring paper title keywords in the project configuration file
• Automatically call the arXiv API to obtain papers published on the current day that match the configured keywords
• Support using large language models (LLM) to summarize the content of the obtained papers
• Parse and display basic paper metadata including title, author, abstract, PDF link, publication time
• Integrate with the project's existing workflow stably

新增基于 arXiv 官方 API 的相关功能：

• 支持在项目配置文件中配置论文标题关键词
• 自动调用 arXiv API 获取当日发布且匹配配置关键词的相关论文
• 支持使用大模型对获取到的论文内容进行概括总结
• 解析并展示论文标题、作者、摘要、PDF 链接、发布时间等基础元数据
• 稳定接入项目现有业务流程

3. Additional context. 补充说明

arXiv provides a complete public API for developers, which can be used to implement keyword filtering and daily paper retrieval:
arXiv 官方提供完善的公开 API，可实现关键词筛选与当日论文获取，文档地址：
https://info.arxiv.org/help/api/user-manual.html

[Thysrael]

I think this is a very good request. However, I feel there are still some things that need to be thought through more carefully.

In my view, the filtering and enrichment logic for research papers is different from that for news. A paper’s original abstract is already quite dense and comprehensive in terms of information—so what can AI really add here? Would it just be summarizing it again? Do you have any good suggestions?

[Thysrael]

1. Contribution Dissection: Extract structured insights beyond the abstract
2. Credibility Assessment
3. Paper Lineage