Zotero MCP

Zotero MCP integrates Zotero research libraries with AI assistants including Claude, ChatGPT, Cherry Studio, Cursor, and Chorus through the Model Context Protocol. It enables users to search, summarize, and analyze library contents directly within those assistants.

The tool provides AI-powered semantic search that uses vector-based methods to identify conceptually related papers beyond keyword matches. It supports multiple embedding models such as local free options, OpenAI, Gemini, and HuggingFace along with intelligent similarity scoring, an auto-updating database with configurable synchronization, and an optional full-text content indexing feature. Standard search covers papers by title, author, content, or tags with advanced multi-criteria logic and tag filtering. Retrieval functions return detailed metadata, full-text content, attachments, PDF annotations, and notes while supporting creation of new notes and compatibility with Zotero's native annotation system.

Write capabilities allow adding papers by DOI or URL, managing collections, updating metadata, performing batch tag operations, finding and merging duplicates, and exporting citations in BibTeX format. The software includes smart update detection that preserves configurations and optional extras for semantic search or PDF extraction. It operates as a command-line interface installed via the uv package manager with commands such as uv tool install zotero-mcp-server or via pip for Python environments. Configuration supports both local Zotero access and remote Web API connections using environment variables for API keys and library identifiers. Setup commands auto-configure for specific AI clients, and a comprehensive set of API tools covers semantic search, item retrieval, annotation handling, note creation, and library management operations.

Zotero MCP addresses the challenge of connecting reference management systems to AI-driven research workflows.