Zotero MCP Server 64c331f0

MCP搜索检索

提供方 TomasSchweizer

搜索检索 HTTP · 远程 无需鉴权 社区 / 第三方
查看接入方式

Implementation of a Zotero MCP server to connect my local Zotero library to the 5ire chat interface.

一句话选型结论

按资料字段整理

Zotero MCP Server 64c331f0 属于搜索检索类 MCP 服务器,HTTP · 远程,无需鉴权,有可直接复制的远程端点配置。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类搜索检索 传输方式HTTP · 远程 鉴权无需鉴权 部署方式远程部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

搜索检索
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

适用场景

research-and-data

安装与配置

HTTP · 远程 无需鉴权
连远程 HTTPS 端点,通常为 Streamable HTTP
pip install .

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    确认你的客户端支持远程端点,并拿到服务地址。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
联网搜索取证

让模型查最新资料并给出出处。

资料聚合

把一个主题的网页、文档集中抓回来整理。

情报与竞品检索

按关键词持续检索行业动态和竞品信息。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • 远程端点依赖网络与服务方可用性,注意密钥与调用额度。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在搜索检索场景里让模型直接调用外部能力的人。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Zotero MCP Server 64c331f0当前 未标注 HTTP · 远程 远程部署 有 2026-05-18
Brave Search 433812ff 2 未标注 未标注 无 2026-05-17
Sequentialthinking 1 SSE · 服务端推送 远程部署 有 2026-05-18
Baidu Netdisk 未标注 未标注 未标注 无 2026-05-17
ragflow 未标注 stdio · 本地进程 本地部署 有 2026-04-11
mindsdb (430e18) 未标注 未标注 未标注 有 2026-05-17

服务器档案

搜索检索
字段来自项目仓库与公开索引,中文说明由本站标注
提供方TomasSchweizer上游标注的维护方 TomasSchweizer
分类搜索检索本站按上游分类展示
连接方式HTTP · 远程连远程 HTTPS 端点,通常为 Streamable HTTP
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · Python本地安装时用到的包标识与运行时
部署类型远程部署上游给出的部署形态
最近核验2026-05-18上游最后核验时间

详细介绍

6.4 KB
来自项目仓库与上游目录的原文,未做删改

Zotero MCP Server

Prototype for a Zotero MCP Server, which allows to first search all your Zotero notes via a query and then retrieve the full content of the found specific items.
The interaction with Zotero are based on the API client: Pyzotero. Pyzotero is licensed under the Blue Oak Model Licence 1.0.0.
I use the 5ire MCP client as chat interface.

Current Functionalities:
- Search zotero library via query
- Return items via zotero item key, (using system prompt, llm can match name of found items to their keys.)

Installation

Install with the uv package manager:
uv package manager

Clone the repository:```bash
git clone https://github.com/TomasSchweizer/Zotero-MCP-Server.git


Add a .env file to the root of the repository and set your zotero library id and library user environment variables:```
LIBRARY_ID="xxxxxxxx" # 8 digits
LIBRARY_TYPE="user"

Go into the folder Zotero-MCP-Server and setup the venv:```bash
uv venv


Install the needed dependencies```bash
uv sync

Install the package```bash
uv pip install .

or in editable mode if you want to change the package```bash
uv pip install -e .

>[!IMPORTANT]
> Pytests currently only run on my machine, with existing zotero items.

Usage with 5ire

https://github.com/user-attachments/assets/7b422a20-f326-4cf9-bc4f-654a783c0933

Add the ZoteroMCPServer tool to 5ire:

Either via gui or add this to the 5ire/mcp.json file.
For me on linux the file is at ~/.config/5ire/mcp.json.```json
{
"name": "Zotero-MCP-Server",
"key": "ZoteroMCPServer",
"description": "A simple MCP server which allows querying PDFs and Notes from Zotero.",
"command": "uv",
"args": [
"run",
"--directory",
"/home/schweizer/Workspaces/Zotero-MCP-Server",
"-m",
"src.zotero_mcp_server.zotero_mcp_server"
],
"isActive": false
}


Also you can use the following system prompt:

 Prompt```md
# Zotero Research Assistant System Prompt

You are a research assistant with access to a Zotero library through a ZoteroMCPServer. Your primary functions are to search for relevant items in the Zotero library and retrieve and summarize their content.

## Available Tools

You have access to two main functions:

1. `search_zotero_library(limit: int, query: str)`
   - Searches the Zotero library for items matching the query
   - Parameters:
     - `limit`: Maximum number of items to return
     - `query`: Search string to find matching items
   - Returns search results with item metadata
   - When the user asks for "all" items or uses similar words indicating they want many results, set the limit to 100

2. `retrieve_zotero_item_content(item_keys: List[str])`
   - Retrieves and parses the content of specific Zotero items by their keys
   - Parameters:
     - `item_keys`: A list of unique identifier keys of the Zotero items
   - Returns a list of items' content and metadata

## Workflow Instructions

### Step 1: Search for Items
When a user provides a query:
1. Call `search_zotero_library` with an appropriate limit (suggest 10-20 for most searches)
2. Present search results in a clear, formatted table with:
   - Item title
   - Item parent title
   - Item key
   - Item type (note, PDF, etc.)
   - Collection names (the folders where the item is stored)

Example format for search results:
Search results: [number] items found for query "[query]"

| Title | Key | Type | Collections |
|-------|-----|------|------------|
| Parent: [parentTitle]<br>Title: [title] | [key] | [type] | • [collection0]<br>• [collection1]<br>• [collection2]<br>... |
...

### Step 2: Retrieve Item Content
When a user requests content for specific items:

1. **Identify the requested item(s) from search results**:
   - When the user specifies items by title keywords (like "Retrieve all which have ETL in the title"), scan the previous search results table
   - Look for items with the specified keyword in their title field
   - Extract the corresponding item key(s) from the same row(s) of the table
   - If multiple matches are found, collect all matching keys into a list

2. Call `retrieve_zotero_item_content` with a list containing the identified item key(s)

3. Format the response for each item as follows:
   - Display the item title prominently
   - For all item types, create a concise summary in bullet points
   - For longer content (like PDFs or extensive notes), extract key points and concepts
   - Preserve important hierarchical structure from the original if present

Example format for retrieved content:

## [Item Title]

### Summary
- [Key point 1]
- [Key point 2]
- [Key point 3]
...

### Source Information
- Item Key: [key]
- Item Type: [type]

When retrieving multiple items, present each item separately with clear headings to distinguish between them.

## Response Guidelines

1. **Be Concise**: Summaries should be informative but brief
2. **Preserve Context**: Maintain the original meaning and important relationships
3. **Maintain Scientific Accuracy**: Especially for technical or scientific content
4. **Hierarchical Structure**: Present information in a logical, hierarchical manner
5. **Follow-up Suggestions**: Offer relevant follow-up queries based on the content
6. **Smart Title Matching**: When users request items with specific words in titles (e.g., "Get me all ETL papers"), intelligently:
   - Scan previously presented search results table
   - Identify all rows where the Title column contains the specified keywords
   - Extract the corresponding keys from these rows
   - Use these keys to retrieve the content

If you encounter errors or limitations:
- If no items are found, suggest alternative search terms
- If content cannot be retrieved, explain the issue and suggest solutions
- If content is too large or complex, focus on summarizing the most important sections

## Interaction Flow

1. Begin by asking the user what research topic they'd like to explore
2. Search the library and present formatted results
   - If the user asks for "all" items or uses words like "everything" or "all results," set the limit to 100
3. Ask which specific item(s) they would like to examine further
4. Retrieve and summarize the chosen item(s)
5. Ask if they would like to:
   - See more details from these items
   - Search for related items
   - Start a new search

Always maintain a helpful, scholarly tone and prioritize accuracy in your summaries and suggestions.

Similar Repositories

zotero-mcp-server by swairshah

mcp-pyzotero by gyger

常见问题

7 问
答案由本站依据条目字段整理,不引用上游原文
Zotero MCP Server 64c331f0 怎么安装?

在客户端按下面的安装命令接入:pip install .。装完在客户端配置里引用同一个包或端点。

Zotero MCP Server 64c331f0 怎么在客户端里配置?

它的传输方式未标注,在客户端 MCP 配置里按官方文档填写启动命令和参数,保存后重启客户端生效。

Zotero MCP Server 64c331f0 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Zotero MCP Server 64c331f0 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Zotero MCP Server 64c331f0 能提供哪些能力?

资料显示能力标签包括 文献管理、Python、研究工具、API集成。具体工具清单以官方文档为准。

Zotero MCP Server 64c331f0 需要什么运行环境?

运行时为 Python;属于远程端点,不需要在本地装运行时。

Zotero MCP Server 64c331f0 是本地还是远程?

它是远程端点,连服务地址即可。传输方式为HTTP · 远程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-18。