Zotero MCP

Read-only MCP server for your local Zotero library. Browse collections, inspect paper metadata, and extract full-text from PDFs -- all via FastMCP tools.

Requirements

• Zotero desktop synced to ~/Zotero/zotero.sqlite
• Python 3.13+
• uv
• pdftotext (for paper_text) -- usually from poppler-utils

Setup

uv sync

Tools

list_library

List all collections and papers as a formatted tree.

uv run python src/main.py

Argument	Type	Description
query	string (optional)	Filter by collection name, paper title, or author (case-insensitive)

Examples:

# Full library
await client.call_tool("list_library", {})
# └── Bachelorarbeit (42 papers)
#     │ [book] Pro Spring Boot 3 (2024) - Felipe Gutierrez
#     ├── GraalVM (4 papers)

# Filtered
await client.call_tool("list_library", {"query": "serverless"})
# └── Serverless (8 papers)
#     [journalArticle] The rise of serverless computing (2019) - Unknown

paper_details

Get full metadata for a paper by its item key.

Argument	Type	Description
item_key	string (required)	Zotero item key (shown in list_library output)

Example:

await client.call_tool("paper_details", {"item_key": "T3ZCDWC7"})
# Title: MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark
# Type: preprint
# Authors: Yubo Wang, Xueguang Ma, ...
# date: 2024-11-06
# DOI: 10.48550/arXiv.2406.01574
# url: http://arxiv.org/abs/2406.01574
# abstractNote: In the age of large-scale language models...
# Collections: Studienarbeit
# PDF: Wang et al. - 2024 - MMLU-Pro A More Robust ....pdf

Returns: title, type, key, dates, authors, all field metadata, collection memberships, and PDF attachment info.

paper_text

Extract the full text of a paper's PDF using pdftotext.

Argument	Type	Description
item_key	string (required)	Zotero item key

Example:

await client.call_tool("paper_text", {"item_key": "T3ZCDWC7"})
# MMLU-Pro: A More Robust and Challenging
# Multi-Task Language Understanding Benchmark
#
# Yubo Wang*, Xueguang Ma*, Ge Zhang, ...
# ...

Usage

# stdio transport (default for MCP clients)
uv run python src/main.py

# fastmcp CLI
uv run fastmcp run src/main.py:mcp

File structure

src/
  main.py    # FastMCP server, tool definitions
  zotero.py  # SQLite queries, data models, formatting

The database is read directly from ~/Zotero/zotero.sqlite. PDFs are resolved from ~/Zotero/storage/. No API key needed.