54yyyu/zotero-mcp

Project overview

It accommodates local offline read access alongside cloud-library web API write operations, and offers an optional terminal CLI for library management without requiring a coding background.

Project type

MCP · RAG · AI Search

Use cases

Knowledge Q&A · Search & Research

Deployment

Refer to project documentation

License

MIT

Best for

• Researchers who need to access, search, and manage their Zotero libraries through AI assistants or a standalone terminal CLI.
• Users requiring local read access without an API key, combined with web API write operations.

Key capabilities

• Finds papers, articles, and books by title, author, or content, providing multi-criteria and semantic search options.
• Retrieves item metadata in Markdown or BibTeX format and provides full text when available in the Zotero library.
• Extracts and searches PDF annotations, accesses native Zotero annotations, and creates or updates notes and annotations.
• Adds papers and files, manages collections and metadata, batch-updates tags, and previews duplicate merges.
• Provides vector-based similarity search over the research library with selectable embedding models and similarity scores.
• Lets users search, browse, and edit a Zotero library from a terminal without needing an AI assistant.
• Optionally reports citation tallies and scans the library for retracted or corrected papers.

Limitations and risks

• The Zotero local API supports read operations only; write operations route through the Zotero web API.
• Vector-based semantic search is not part of the base installation and requires installing the optional semantic extra, which includes ChromaDB.
• Full text retrieval functions only when the full text resource is already present in the Zotero library.
• The optional semantic search feature supports remote inference using OpenAI, Gemini, or Ollama API embedding models. Local orchestration does not preclude remote API calls for this feature.

Getting started

• Setup is considered medium difficulty because the documented requirements include Python, Zotero, and an MCP-compatible client. Installation is handled through Python package commands and server setup commands.

Alternatives and comparisons

• Provides question-answering over local documents using hybrid search and re-ranking, serving as an alternative for searching local text files rather than a connected Zotero library.
• Provides fulltext search and retrieval over crawled web content via MCP, which may apply to users seeking search over web archives rather than structured research papers.
• Provides local desktop search using text, image, and voice inputs for users who need to search across multimedia files instead of a Zotero library.

Project comparisons

• 54yyyu/zotero-mcp vs Advanced Local RAG System
• 54yyyu/zotero-mcp vs Web Crawler Search Interface
• 54yyyu/zotero-mcp vs Multimodal Local File Search

Evidence and sources

• README: The Zotero local API must be enabled for the MCP server to work.
• GitHub project description: Zotero MCP: Connects your Zotero research library with Claude and other AI assistants via the Model Context Protocol to discuss papers, get summaries, analyze citations, and more.
• README: - Find papers, articles, and books by title, author, or content - Perform complex searches with multiple criteria - Browse collections, tags, and recent additions - Semantic searc…
• README: - Retrieve detailed metadata for any item (markdown or BibTeX export) - Get full text content (when available) - Look up items by BetterBibTeX citation key
• README: - Extract and search PDF annotations with page numbers - Access Zotero's native annotations - Create and update notes and annotations - Extract PDF table of contents / outlines (r…

Reviewed at: 2026-07-16T07:16:09Z

Open original GitHub project