LLM For Zotero is a free, open-sourced Zotero add-on developed by Yilewang, enhancing the Zotero PDF reader with Large Language Model capabilities. Distributed as a .xpi file, this MCP server tool integrates AI directly into research workflows. It targets academics, researchers, and PhD students, providing advanced functionalities for literature management and knowledge extraction. The program acts as a research agent system, bringing sophisticated AI assistance to Zotero's environment.

What does LLM For Zotero connect to?

LLM For Zotero primarily integrates with the Zotero PDF reader interface, acting as a Zotero plugin. It connects to various AI clients and services through streamable HTTP for API providers like OpenAI, Anthropic, and Gemini. For WebChat modes, it utilizes a browser extension, Sync for Zotero, connecting to services such as ChatGPT. Additionally, it can connect to local OpenAI-compatible models like Ollama and LM Studio, and runs through local runtimes for Codex App Server and Claude Code Mode, extending its reach across diverse AI ecosystems.

Core features and tools

LLM For Zotero exposes several callable tools. Users can Chat with the current PDF, Get grounded answers with citations, and Summarize a full paper or selected text. It enables Agent Mode for library-wide operations, including searching, tagging, and metadata updates. The tool supports Filesystem and scripting for local file read/write and Zotero JavaScript execution, with safety features like undoing write actions. Other capabilities include Save answers to Zotero notes, Create books and other Zotero item types, Update different metadata fields across multiple items, and Perform built-in web and academic literature research.

Who is it for?

LLM For Zotero is designed for researchers, academics, and PhD students who rely on Zotero for their literature management. It provides advanced functionalities for interacting with research papers, automating organizational tasks, and generating notes. The tool is particularly useful for those seeking to integrate AI directly into their reading and library management processes to enhance productivity and knowledge extraction. Skip this if you do not use Zotero for managing your research literature.

How to get started

Access is granted through an API key for most providers, pasted into Zotero Preferences under the llm-for-zotero tab. Alternatively, WebChat mode works without an API key by installing the Sync for Zotero browser extension.

Use cases

• Interactive Paper Chat and Summarization: Users can ask questions about open PDFs, selected text, or figures, receive grounded answers with citations, and summarize papers or specific sections directly within the Zotero reader.
• Automated Library Management and Organization: Through Agent Mode, the tool can perform library-wide tasks such as searching, tagging, updating metadata, organizing collections, and importing new papers via identifiers.
• Integrated Note-taking and Knowledge Base Sync: It allows users to generate and save research notes in Markdown format to local folders like Obsidian or Logseq, complete with metadata, citations, and extracted figures, or directly to Zotero item notes.

LLM For Zotero offers a valuable set of AI-powered features for Zotero users, enhancing document interaction and library management. Its integration with multiple AI providers and flexible authentication models make it adaptable for various research needs. The tool's ability to automate tasks and provide grounded answers with citations can significantly aid academic workflows.