Zotero MCP
Zotero MCP
面向本地 Zotero 库的只读 MCP 服务器。可浏览分类、查看论文元数据,并通过 FastMCP 工具从 PDF 中提取全文。
要求
Zotero 桌面版已同步到
~/Zotero/zotero.sqlite(Linux 默认路径)Python 3.13+
Related MCP server: zotero-mcp
配置
这是一个 MCP 服务器。你需要在编码代理的 MCP 配置中注册它,之后代理即可使用其工具。
大多数代理接受类似的配置。例如,在 Opencode 中,你将其添加到 opencode.json:
{
"mcp": {
"zotero-mcp": {
"command": [
"uvx",
"--from",
"git+https://github.com/404Simon/zotero-mcp",
"zotero-mcp"
],
"enabled": true,
"type": "local"
}
}
}完成此设置后,代理会自动发现工具,你可以直接向它提问,例如:
“我的库中有哪些关于 RAG 的论文?最新一篇的摘要是什么?”
工具
list_library
以格式化树形结构列出所有分类和论文。可选按分类名称、论文标题或作者过滤。
参数 | 类型 | 描述 |
|
| 按分类名称、论文标题或作者过滤(不区分大小写) |
每篇论文行都包含其条目键,显示在 [KEY] 方括号中。可将该键用于 paper_details 或 paper_text。过滤后的输出仅显示匹配的子树,包括匹配的分类及其祖先。
├── AI (2 papers)
│ [N3G6XKB9] [preprint] Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (2021) - Patrick Lewis
│ [PPJJCMXJ] [book] Grundkurs Künstliche Intelligenz: eine praxisorientierte Einführung (2021) - Wolfgang Ertel
├── Bachelorarbeit (42 papers)
│ ├── GraalVM (4 papers)
│ │ [DSUERN67] [book] Supercharge your applications with GraalVM ... (2021) - A. B. Vijay Kumar
│ └── Java Performance (1 papers)
│ [Q3ANMGC3] [conferencePaper] Applying Optimizations for Dynamically-typed Languages to Java (2017) - Matthias Grimmer
│ [MTSF327R] [book] Pro Spring Boot 3: An Authoritative Guide with Best Practices (2024) - Felipe Gutierrez
├── Studienarbeit (49 papers)
│ [T3ZCDWC7] [preprint] MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark (2024) - Yubo Wang
├── T3000 (4 papers)
│ [DTZF77Y4] [webpage] Conventional Commits (n.d.) - Unknown
├── TheGreenEpoch (8 papers)
│ [YINRQ63P] [preprint] Distributed LLM Pretraining During Renewable Curtailment Windows (2026) - Philipp Wiesner
└── VesSkel (19 papers)
[ZUXQUGHW] [journalArticle] Open-source analysis and visualization of segmented vasculature datasets with VesselVio (2022) - Jacob R. Bumgarnerpaper_details
通过条目键获取论文的完整元数据。可从 list_library 输出(显示为 [KEY])或 search_papers 结果中获取该键。
参数 | 类型 | 描述 |
|
| Zotero 条目键(在 list_library 输出或 search_papers 结果中显示为 |
返回标题、类型、键、添加/修改日期、作者、所有字段元数据、所属分类以及 PDF 附件信息:
Title: MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark
Type: preprint
Key: T3ZCDWC7
Authors: Yubo Wang, Xueguang Ma, Ge Zhang, Yuansheng Ni, ...
date: 2024-11-06
DOI: 10.48550/arXiv.2406.01574
url: http://arxiv.org/abs/2406.01574
abstractNote: In the age of large-scale language models, benchmarks like the Massive Multitask Language Understanding (MMLU) ...
Collections: Studienarbeit
PDF: Wang et al. - 2024 - MMLU-Pro A More Robust and Challenging Multi-Task Language Understanding Benchmark.pdfsearch_papers
按标题或作者搜索所有论文。返回包含条目键的结构化结果。
参数 | 类型 | 描述 |
|
| 搜索词(不区分大小写,匹配标题和作者) |
[{"key": "ZUXQUGHW",
"title": "Open-source analysis and visualization of segmented vasculature datasets with VesselVio",
"type": "journalArticle", "year": "2022",
"first_author": "Jacob R. Bumgarner",
"authors": ["Jacob R. Bumgarner", "Randy J. Nelson"],
"url": "https://linkinghub.elsevier.com/retrieve/pii/S2667237522000443"},
{"key": "6DU4XPQQ",
"title": "Robust Vessel Segmentation in Fundus Images",
"type": "journalArticle", "year": "2013",
"first_author": "A. Budai",
"authors": ["A. Budai", "R. Bock", "A. Maier", "J. Hornegger", "G. Michelson"],
"url": "http://www.hindawi.com/journals/ijbi/2013/154860/"}]paper_text
使用 PyMuPDF(作为 Python 依赖打包,无需系统工具)提取论文 PDF 的全文。需要 PDF 附件存储在 ~/Zotero/storage/ 中。
参数 | 类型 | 描述 |
|
| Zotero 条目键(在 list_library 输出或 search_papers 结果中显示为 |
返回原始提取文本,以标题、作者和摘要开头:
MMLU-Pro: A More Robust and Challenging
Multi-Task Language Understanding Benchmark
1Yubo Wang∗, 1Xueguang Ma∗, 1Ge Zhang, 1Yuansheng Ni, 1Abhranil Chandra, ...
1University of Waterloo, 2University of Toronto, 3Carnegie Mellon University
Abstract
In the age of large-scale language models, benchmarks like the Massive Multitask
Language Understanding (MMLU) have been pivotal in pushing the boundaries
of what AI can achieve in language comprehension and reasoning across diverse
domains. ...文件结构
src/
main.py # FastMCP server, tool definitions
zotero.py # SQLite queries, data models, formatting数据库直接从 ~/Zotero/zotero.sqlite 读取。PDF 从 ~/Zotero/storage/ 解析。无需 API 密钥。
Available Tools
4 toolslist_libraryA
List collections and papers in the Zotero library. Optionally filter by collection name, paper title, or author. Each paper line includes its item key in [KEY] brackets — use that key in paper_details or paper_text.
Inputs: query. Returns: result
paper_detailsA
Get full metadata for a paper by its item key. Obtain the item_key from list_library output (shown as [KEY] before each paper) or from search_papers results.
Inputs: item_key (required). Returns: result
paper_textA
Extract full text from a paper's PDF using pdftotext. Obtain the item_key from list_library output (shown as [KEY] before each paper) or from search_papers results.
Inputs: item_key (required). Returns: result
search_papersA
Search papers by title or author and return structured results with item keys. Each result includes key, title, type, year, first_author, authors, and url. Use the key in paper_details or paper_text to get full metadata or PDF text.
Inputs: query (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
The tools are mostly distinct: list_library for browsing, search_papers for targeted search, paper_details for metadata, and paper_text for PDF content. However, list_library and search_papers both filter by title/author, which could cause some confusion.
The naming mixes verb_noun patterns (list_library, search_papers) with noun_noun patterns (paper_details, paper_text). While the paper_* prefix helps, the overall convention is not fully consistent.
With only 4 tools, the server is well-scoped for its purpose of accessing a Zotero library. Each tool is essential and there is no unnecessary bloat.
The tool set covers the core read workflow: discovering papers (list/search), retrieving full metadata, and extracting PDF text. Missing write operations and a dedicated collection endpoint are minor gaps that agents can work around.
Maintenance
Related MCP Connectors
Remote MCP server for full read/write access to a Zotero library
Zotero MCP server for Claude and ChatGPT: search, citations, safe writes, PDF passages and pages.
Read-only MCP server exposing a user ORANO library to their own AI agent.
Personal knowledge base MCP server with semantic search, auto-categorization, metadata extraction
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceRead-only MCP server for browsing, searching, and exporting a Zotero library from AI assistants.-
- AlicenseAqualityDmaintenanceRead-only MCP server that lets Claude or any MCP client search and retrieve metadata, notes, full text, citations, and BibTeX from your local Zotero library via its built-in API.11MIT
- AlicenseNot gradedqualityBmaintenanceLocal read-only MCP server for Zotero libraries, enabling search, retrieval, and full-text access via the Zotero Web API.MIT
- AlicenseAqualityCmaintenanceMinimal Zotero MCP server that enables searching your Zotero library, discovering papers across eight sources, downloading PDFs via a multi-level cascade, and importing them into Zotero.61MIT