academia_mcp

概述

Academia MCP 服务器是基于 Python 的综合性工具，专门设计用于自动化科研工作流。它集成了多个学术平台，包括 ArXiv、ACL Anthology、Hugging Face 和 Semantic Scholar，以及搜索引擎。服务器提供论文搜索、PDF 下载和阅读、LaTeX 编译等功能，并支持可选的 LLM 驱动的文档问答和研究提案生成。它支持多种传输协议（stdio、sse、streamable-http），并包含基于令牌的身份验证功能，可在生产环境中安全部署。

试试问 AI

装完之后，这里有 6 个你可以让 AI 做的事：

• arxiv_search

  Query arXiv with field-specific queries and filters.

  试试:Find recent machine learning papers about transformers试试:Search for papers in computer vision from 2022 to 2023
• arxiv_download

  Fetch a paper by ID and convert to structured text.

  试试:Download paper 2310.05217 from arXiv试试:Get HTML version of paper 1807.03843
• anthology_search

  Search ACL Anthology with fielded queries and date filtering.

  试试:Find papers about NLP in ACL Anthology试试:Search for machine learning papers from 2020
• hf_datasets_search

  Find Hugging Face datasets with filters and sorting.

  试试:Search for datasets about natural language processing试试:Find computer vision datasets with more than 1000 samples
• s2_get_citations

  List papers citing a given arXiv paper from Semantic Scholar.

  试试:Find papers that cite 2306.11646试试:Get citations for paper 1810.04805
• s2_get_references

  List papers referenced by a given arXiv paper.

  试试:Show references for paper 2306.11646试试:Find references of paper 1810.04805
• visit_webpage

  Fetch and normalize a web page.

  试试:Visit https://arxiv.org/abs/2306.11646 and summarize试试:Extract content from https://aclanthology.org/
• web_search

  Unified search wrapper using configured search providers.

  试试:Search for recent developments in quantum computing试试:Find papers about transformer architectures
• compile_latex

  Compile LaTeX to PDF in WORKSPACE_DIR.

  试试:Compile my paper.tex to PDF试试:Build research document from LaTeX source
• read_pdf

  Extract text per page from a PDF.

  试试:Read content of paper.pdf试试:Extract text from research_document.pdf
• document_qa

  Answer questions over provided document chunks using LLM.

  试试:What are the main findings in this document?试试:Summarize the key points of this paper
• generate_research_proposals

  Generate research proposals using LLM.

  试试:Generate a research proposal for neural architecture search试试:Create a proposal for studying climate change impacts

可对比工具

安装

pip3 install academia-mcp

uv venv .venv
make install

python -m academia_mcp --transport streamable-http

python -m academia_mcp --transport stdio

{
  "mcpServers": {
    "academia": {
      "command": "python3",
      "args": [
        "-m",
        "academia_mcp",
        "--transport",
        "stdio"
      ]
    }
  }
}

FAQ

这个服务器支持哪些学术数据库？

服务器支持 ArXiv、ACL Anthology、Hugging Face 数据集以及 Semantic Scholar 引用和参考文献。

如何启用文档问答等 LLM 驱动的功能？

设置 OPENROUTER_API_KEY 环境变量，以启用 document_qa 和研究提案助手等基于 LLM 的工具。

是否支持 LaTeX 编译？

是的，服务器包含 LaTeX 编译和模板工具，但需要在系统上安装 LaTeX 发行版。

academia_mcp 对比