AI搜索如何工作
用自然语言提问
语义搜索理解您的本意，而不只是字面词语。像问同事一样提问即可，无需先猜对关键词。 示例：“哪些干预措施能降低心力衰竭患者的30天再入院率？”
检索并重新排序
Paperguide 先在文献中全面检索，再按相关性和质量——SJR 分区、SNIP、SJR 分值和引用指标——重新排序，突出最有力的论文。
生成有引用支撑的答案
从排名前20的论文中综合出答案，每条结论都链接到其来源，发现和证据一并呈现。
不止于答案
让证据始终触手可及
从一个问题走向下一步所需的一切。
- 每个答案背后都附有一份排好序的参考文献列表，可直达论文本身。
- 追问会延续上下文，逐步细化而不必从头再来。
- 结果可直接保存到您的 Paperguide 参考文献管理器。
数据库
基于2亿+篇同行评审论文
答案反映文献的全貌，而非一角，取自各大开放科学数据源。
- PubMed 生物医学与生命科学文献。
- arXiv 物理、数学与计算机科学预印本。
- OpenAlex 覆盖所有学科的开放学术著作目录。
- Semantic Scholar 跨学科语料库，配有 AI 提取的元数据。
还有您自己的论文，可从 Zotero、Mendeley 或您的 Paperguide 参考文献管理器导入。
一个带引用的答案，为您省下一个下午
带引用的答案，快速可得
问一句当前研究怎么说，几分钟内得到带引用的答案，别人追问出处时随时可以引用。
综述之前先摸底
在投入完整文献综述之前，先看清某个问题有哪些证据、证据有多强。
回应审稿意见
当审稿人要求补充近期研究时，搜一下，看清现状，用引用作答。
只看最有力的论文
按期刊和引用信号排序，突出最相关、最高质量的研究，无需在其余论文中跋涉。
为每一类研究团队提供带引用的答案
找到您的工作流。每张卡片都链接到该类团队使用 Paperguide 的方式。
In practice
96%less time on literature reviews
Cited clinical evidence in minutes under assessment deadlines, with every claim traceable.Read case study
Multilingualclinical synthesis, in French
Clinical literature across languages synthesized into one cited answer, in a hospital setting.Read case study
2-3×more research areas monitored
Supervising more theses across more fields, with reviews that stay structured and cited.Read case study
Daysnot weeks, for a rapid assessment
Rapid evidence assessments where the trail is a byproduct of the work, one click from any claim.Read case study
研究者怎么说
Paperguide search, literature review and writer tools have made my research so much easier. I can complete my whole research in Paperguide, without even switching to Google Scholar and ChatGPT.
Gibbs H.Researcher
I ask a question and get an answer with the papers attached. The citations are what make it usable for real work, not just a starting point.
Nadia H.PhD Candidate
I ask a question the way I would ask a colleague and get a cited answer from the actual literature. The quality signals on each result mean I trust the top papers instead of second-guessing them.
Pedro G.Research Scientist
It scoped a topic for me in minutes and showed how strong the evidence was before I committed to a full review. That saved me from a dead-end search.
Aditi R.PhD Candidate