Zotero MCP by developer Cookjohn is an MCP (Model Context Protocol) server designed to bridge Zotero research libraries with AI assistants. Built using TypeScript, this tool is distributed as a Zotero plugin, specifically an .xpi file, for direct installation. It enhances research workflows by enabling AI interaction with Zotero content, allowing for advanced search and content extraction capabilities. Researchers and knowledge workers benefit from its ability to connect Zotero's reference management system with various AI clients, expanding the utility of their collected resources.

What does Zotero MCP connect to?

Zotero MCP connects directly to the Zotero reference management system. It can interface either through Zotero's local desktop application API or its Zotero Web API. This tool functions as an MCP (Model Context Protocol) server, facilitating communication between Zotero libraries and various AI assistants. This integration allows AI clients to access and process information stored within a user's Zotero collection, enhancing research and knowledge management tasks.

Core features and tools

Zotero MCP provides several capabilities for interacting with Zotero libraries. These include Smart Search, which offers multi-dimensional library search across title, creator, year, tags, fulltext, and semantic data, complete with boolean operators and relevance scoring. It also offers Content Extraction to retrieve PDF full-text, notes, abstracts, and webpage snapshots, with fine-grained mode control. Additionally, it supports Annotation Analysis, allowing users to search and analyze PDF highlights and annotations by color, tags, and keywords.

Who is it for?

Zotero MCP is intended for researchers and knowledge workers who aim to integrate their Zotero libraries with AI assistants. This includes individuals who need to perform advanced searches, extract detailed content, or analyze annotations using AI-powered tools. Its design supports improved interaction with research materials for deeper analysis and summary generation. Skip this if you do not use Zotero for managing your research library.

How to get started

Access is granted through enabling the local Zotero API. AI clients connect via a Streamable HTTP URL, such as http://127.0.0.1:23120/mcp, using the Streamable HTTP transport protocol.

Use cases

• Performing multi-dimensional searches within a Zotero library using natural language queries from an AI assistant.
• Extracting full text, metadata, and notes from Zotero items for AI-powered summarization and analysis.
• Analyzing PDF highlights and annotations, and potentially creating new notes or annotations via AI interaction.

Zotero MCP offers a powerful bridge between Zotero reference management and AI assistants, enabling advanced library interaction. It allows researchers to use AI for enhanced search, content extraction, and annotation analysis directly within their Zotero collections. The plugin-based distribution simplifies its integration into existing Zotero setups.