Zotero MCP integrates your Zotero research library with various AI assistants like ChatGPT and Claude, enabling search, reading, citing, and annotating papers directly within AI agents or a Zotero chat panel. It streamlines academic workflows by providing AI agents access to your full research collection through the Model Context Protocol.

zotero-mcp

Active·★ 5.5k·MIT·Updated 2026-10-16

★ Essential★ Trending

Zotero MCP connects a Zotero research library to AI assistants through the Model Context Protocol. The README says one package, zotero-mcp-server, ships three ways in: an MCP server for chat apps without a shell (Claude Desktop, ChatGPT, Cherry Studio, Chorus), a zotero-cli with an agent skill for agents that have a shell (Claude Code, Codex, Cursor, Windsurf, Gemini CLI and others), and a Zotero Agent chat panel inside Zotero that knows which item, page and selection you have open. It states that the CLI and skill route keeps about 98 tokens in context until needed, against roughly 13k for the full MCP tool list.

The feature list covers search by title, author, tag, collection, full text or meaning (semantic search with local, OpenAI, Gemini or Ollama embeddings), reading metadata, BibTeX, full text and PDF page ranges with page images where text extraction garbles maths or figures, annotations placed on exact words, figures, tables or equations, and writes such as adding papers by DOI, URL, ISBN, BibTeX or file, managing collections and tags and merging duplicates. Scite citation tallies and retraction alerts are optional.

Install with `uv tool install zotero-mcp-server` or pip on Python 3.10+, then enable Zotero's local API under Settings, Advanced in Zotero 7+. In local mode reads come straight from zotero.sqlite. For writes, Zotero 10+ needs a one-time `zotero-mcp authorize-local`; older Zotero can write through the web API with ZOTERO_API_KEY and ZOTERO_LIBRARY_ID. Heavier features such as semantic search are optional extras. It is MIT licensed and the last push was 2026-10-15. This review is based on the README and repository data; we have not run it ourselves.

#Zotero#AI Agent#Research Library#LLM#Academic Tools#Semantic Search#PDF Annotation#Context Protocol

Features

01Search Zotero library (including semantic search) by title, author, tag, collection, or full text.

02Read metadata, BibTeX, full text, and page ranges of PDFs, with page images for complex content.

03Annotate PDFs with highlights, area boxes, and notes, and extract existing PDF annotations.

04Manage Zotero items: add papers by DOI/URL/ISBN/BibTeX/file, organize collections and tags, merge duplicates.

05Flexible data access: local reads from zotero.sqlite or writes through Zotero local API / web API.

Why choose it

+Three routes in one package: MCP server, CLI with agent skill, and a chat panel inside Zotero

+Searches by metadata, full text or meaning, and reads PDF page ranges with page images for maths, figures and tables

+Writes as well as reads: add papers by DOI, URL, ISBN or BibTeX, manage tags and collections, merge duplicates, place annotations on exact text

+In local mode reads come straight from zotero.sqlite; MIT licensed with a recent push (2026-10-15)

Trade-offs

!Write access can add, merge and change items in your library, so approve writes deliberately and keep a backup of your library

!Semantic search and Scite are optional extras, and embeddings can use OpenAI or Gemini, which sends text to those providers unless you choose a local model

!Setup has several steps: install the package, enable Zotero's local API, connect an assistant, and authorise writes

!Aimed at Zotero users; the README reports no results for other reference managers

Maintenance check

ActiveChecked: 2026-10-20

Last commit: 2026-10-16
Commits, last 90 days: 100+
Releases, last 90 days: 21
Latest release: v0.15.1 (2026-10-16)
Open issues: 57
Recent issues with at least one comment: 9/14

Commits per week, last 12 weeks

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

#700 Zotero Agent plugin: all CLI detection fails on Windows (`/bin/sh` probe bypasses the existing `pathSearch`)6 comments · 2026-10-09
#377 Feature request: rename an attachment's filename on disk, not just its title5 comments · 2026-07-11
#493 Feature request: dedupe cross-library copies in semantic search results (with over-fetch so limit stays meaningful)4 comments · 2026-08-24
#604 `update-db` without `--fulltext` indexes every annotation, the SQLite path does not, and the mismatch never self-heals3 comments · 2026-09-25
#591 Attachment file helpers don't know which library they serve — group uploads go to the personal WebDAV3 comments · 2026-09-24

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

Compatibility

Python

Runtime

Verified via docs

Zotero

Application

Verified via docs

AI Agents

Integrates with

Verified via docs

Docker

Deployment

Verified via docs

Quick start

$ uv tool install zotero-mcp-server

Use cases

↳Interacting with your personal Zotero research library using AI chatbots like Claude or ChatGPT.

↳Conducting semantic searches across your papers to find research by meaning, not just keywords.

↳Getting AI agents to explain complex research papers with page citations directly within Zotero.

↳Automating tasks like summarizing key findings or highlighting main claims in PDF documents.

Alternatives

⚡Top VS:

VSragflow★92.6k→VSAgentGuide★11.1k→VSAIGC-Interview-Book★5.0k→VSGPT Researcher★30.2k→

ragflow★ 92.6k

RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs

Features

Why choose it

Trade-offs

Maintenance check

Most discussed open issues

Compatibility

Quick start

Use cases

Alternatives

Comments

Features

Why choose it

Trade-offs

Maintenance check

Most discussed open issues

Compatibility

Quick start

Use cases

Alternatives

Comments