MCP Servers · 54yyyu

zotero-mcp

Zotero MCP is a Python-based integration layer that connects Zotero research libraries with Claude, ChatGPT, and other AI assistants via the Model Context Protocol. It enables AI-assisted discovery, summarization, and citation analysis of academic papers, with optional semantic search using embeddings from local or cloud providers.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/54yyyu/zotero-mcp

4.2k

GitHub stars

351

Forks

Python

Primary language

MIT

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	54yyyu/zotero-mcp
Owner	54yyyu
Primary language	Python
License	MIT — OSI-approved
Stars	4.2k
Forks	351
Open issues	85
Latest release	v0.6.1 (2026-07-03)
Last updated	2026-07-05
Source	https://github.com/54yyyu/zotero-mcp

What zotero-mcp is

A Python MCP server implementing the Model Context Protocol to expose Zotero library operations (search, metadata retrieval, annotation extraction, write operations) and optional ChromaDB-backed semantic search with pluggable embedding backends (default, OpenAI, Gemini, Ollama). Supports local Zotero mode, web API mode, and hybrid mode with configurable batch processing for embeddings.

Quickstart

Get the zotero-mcp source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/54yyyu/zotero-mcp.git2cd zotero-mcp3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Research Literature Discovery & Synthesis

Academics and researchers querying their Zotero libraries conversationally to find relevant papers, extract summaries, analyze citations, and synthesize findings without leaving their AI assistant.

Semantic Literature Search

Finding papers by conceptual similarity (e.g., 'machine learning in neuroscience') rather than keyword matching, using optional embeddings to enable topic-based discovery across large personal libraries.

Annotation & Knowledge Management Augmentation

Extracting, searching, and organizing PDF annotations and notes within AI conversations, enabling interactive review and knowledge consolidation workflows.

Implementation considerations

Choose embedding model early: local (free, privacy-preserving, slower) vs. cloud (better quality, API costs, latency). Switching models later requires rebuilding the semantic index.
Semantic search is optional; base install includes search, metadata, annotations, and writes without ML dependencies. Install semantic/pdf/scite extras only if needed to keep footprint minimal.
Web API mode requires Zotero API key; local mode requires local Zotero installation and sqlite database access. Hybrid mode (read local, write via API) is available for mixed workflows.
PDF outline extraction and EPUB annotations require optional pdf extra (PyMuPDF dependency). Scite citation intelligence requires separate optional extra.
Batch API for OpenAI embeddings is asynchronous; suitable for large libraries but requires monitoring and import step. Realtime embeddings available as fallback.

When to avoid it — and what to weigh

Large-scale institutional repository access — Designed for personal/small-team Zotero libraries, not enterprise-scale research data warehouses or federated search across multiple institutions.
Offline-only environments with no AI assistant integration — The core value is AI-native interaction; if your workflow is purely local command-line or offline, consider Zotero's native UI or alternatives focused on non-AI workflows.
Privacy-critical deployments without full control — Semantic search with external embeddings (OpenAI, Gemini) sends metadata/content to cloud providers; local embeddings mitigate this but semantic search is optional.
Production systems requiring guaranteed SLA/support — Community-driven open-source project with no commercial support contract; incidents rely on maintainer responsiveness and community issue resolution.

License & commercial use

MIT License (MIT). Permissive open-source license allowing commercial use, modification, and distribution with attribution. No restrictions on bundling or selling derived works.

MIT License permits unrestricted commercial use. However, commercial use of external embedding services (OpenAI, Gemini) incurs their own licensing and cost terms. Scite citation intelligence feature uses public API endpoints (no Scite account required per README, but Scite's terms of service for public API use should be verified). No commercial support or SLA from the project itself.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Low
DEV.co fit	Good
Assessment confidence	High

Security considerations

Local mode has no external network exposure (reads from local zotero.sqlite). Web API mode transmits queries and metadata to Zotero servers (uses HTTP via Zotero's official API). Semantic search with external embeddings (OpenAI, Gemini) transmits paper titles/abstracts/content to cloud providers—review vendor privacy policies. Local embeddings (all-MiniLM-L6-v2, Ollama) keep data on-device. PDF extraction and annotation reading are local operations. No explicit encryption, authentication, or authorization layers described; assumes AI assistant (Claude, ChatGPT) handles user authentication.

Alternatives to consider

Zotero native UI + manual AI queries

No additional tool or integration needed. Trade convenience of conversational search and automated summaries for full control over data and no external dependencies.

Semantic Scholar API + custom MCP implementation

Broader academic search but focused on published papers, not personal library management. Requires building your own MCP server if MCP integration is required.

ReadWise + GPT integration

Highlights and annotations from web articles and PDFs integrated with OpenAI. Different scope (captures web content, not bibliography management), but overlaps on AI-assisted knowledge synthesis.

Software development agency

Build on zotero-mcp with DEV.co software developers

Start connecting your Zotero library to Claude and other AI assistants. Install via pip/uv in minutes, with optional semantic search for advanced discovery.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

papersgpt-for-zotero mcp mcp-atlassian

Related on DEV.co

Explore the category and the services that help you build with it.

MCP Servers AI Application Development API Development Talk to DEV.co

zotero-mcp FAQ

Do I need a Zotero account or API key?

Not for local mode (uses local zotero.sqlite). Web API mode requires a free Zotero account and API key. Hybrid mode reads locally but writes via web API, so API key is recommended for write operations.

Does semantic search work offline?

Yes, if you use the default embedding model (all-MiniLM-L6-v2, runs locally) or Ollama. OpenAI/Gemini embeddings require internet and API keys.

What AI assistants are supported?

Claude Desktop is explicitly supported. README mentions ChatGPT, Cherry Studio, Chorus, and Cursor integration, but exact MCP configuration steps for non-Claude clients are not detailed.

How much do embeddings cost?

Default (free, local). OpenAI: ~$0.02–0.06 per 1M tokens depending on model (realtime) or cheaper via Batch API. Gemini: similar pricing structure. Local Ollama: free (compute-only, your hardware).

Software development & web development with DEV.co

DEV.co helps companies turn open-source tools like zotero-mcp into production software. Our software development services cover the full lifecycle — architecture, web development, integration, and maintenance — delivered by software developers and web developers who ship. Engage our software development agency to implement or customize it for your mcp servers stack.

Integrate AI with Your Research Library Today

Start connecting your Zotero library to Claude and other AI assistants. Install via pip/uv in minutes, with optional semantic search for advanced discovery.

Talk to DEV.co Schedule a Consultation