Zotero MCP Bridge
Enables LLM clients to browse and query your Zotero library through tool calls. Provides access to Zotero-specific functions like listing open tabs, searching items, and browsing collections via a local MCP server running inside Zotero.
README
Zotero MCP Bridge
Zotero MCP Bridge is a Zotero 7 plugin that will expose a local Model Context Protocol (MCP) server so LLM clients can browse and query your Zotero library with tool calls. The plugin is written in TypeScript and will integrate the modelcontextprotocol/typescript-sdk for the MCP implementation.
This repository started from the canonical Zotero plugin template, but the template-only samples and UI chrome have been removed so we can focus on the MCP bridge code.
Project Goals
- Ship a local MCP server that runs inside Zotero and surfaces Zotero-specific tools such as listing open tabs, searching items, and browsing collections.
- Provide a lightweight communication shim so LLMs can attach to the running Zotero MCP server.
- Keep the UI surface minimal—most interaction should happen through MCP tool calls rather than extra Zotero menus.
Repository Layout
src/– TypeScript source for the plugin, including the MCP runtime scaffolding.addon/– Static assets packaged with the XPI (manifest, icons, locale strings).test/– Mocha test suite executed vianpm test.zotero-plugin.config.ts– Build configuration forzotero-plugin-scaffold.
Getting Started
npm install
npm run build # bundles the plugin and runs `tsc --noEmit`
During development you can hot-reload the plugin into Zotero with:
npm run start
Run tests in a scaffolded Zotero instance with:
npm test
Note: the build pipeline expects Zotero 7 and the latest Node.js LTS release. See the official developer guide for environment prerequisites.
Contributing
- Track work with
bd(seeAGENTS.md). - Keep code TypeScript-first with strict linting (
npm run lint:check). - Commit code and
.beads/issues.jsonltogether so issue state mirrors code state.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。