papersgpt-for-zotero
A powerful Zotero AI and MCP plugin with ChatGPT, Gemini 3.7, Claude Fable 5, Claude Opus 5, DeepSeek V4, Grok, OpenRouter, Kimi k3, GLM 5.3, SiliconFlow, GPT-o
What is papersgpt-for-zotero?
A powerful Zotero AI and MCP plugin with ChatGPT, Gemini 3.7, Claude Fable 5, Claude Opus 5, DeepSeek V4, Grok, OpenRouter, Kimi k3, GLM 5.3, SiliconFlow, GPT-oss, Gemma 4, Qwen 3.8
Overview of papersgpt-for-zotero
What it is
PapersGPT is a Zotero extension built in Native C++ that indexes and searches PDFs locally. It supports multiple large language models (LLMs) including ChatGPT, Gemini 3.7, Claude Fable 5, Claude Opus 5, DeepSeek V4, Grok, OpenRouter, Kimi k3, GLM 5.3, SiliconFlow, GPT-oss, Gemma 4, and Qwen 3.8. The extension performs all parsing, indexing, and searching on the user's machine with no cloud dependency.
Key features
- AutoPilot Automation: batch-processes 1,000+ papers overnight and saves insights into Zotero Notes.
- SOTA Model Cowork: switches between multiple cloud LLMs or runs local LLMs via Ollama for an air-gapped experience.
- Zotero MCP Server: integrates the local Zotero knowledge base as a skill for agentic platforms like Claude Code, Cursor, or Windsurf.
- Deep cross-document comparison and global library synthesis across the entire Zotero library.
- Verifiable evidence with click-to-jump citations in every AI response.
- 100% offline local LLM support via integrated open-source models, custom OpenAI-compatible endpoints, or Ollama.
- Zero-byte data leakage: all document processing stays on the device.
Use cases
Research: cross-reference thousands of papers while maintaining 100% data sovereignty. Finance: synthesize decade-long 10-K filings without cloud risk. Legal: perform high-precision discovery across thousands of case files.
Getting started
- Download the latest PapersGPT from the repository. 2. Install the downloaded Zotero plugin file. 3. Open a PDF from your Zotero library and click the PapersGPT button on the top menu of the PDF viewer, or use the keyboard shortcut (macOS: Command + Enter, Windows: Ctrl + Enter). 4. Select an LLM model and configure the API key. 5. Ask questions using built-in prompts or custom input.
License and pricing
License: AGPL-3.0. Pricing: open source.
Based on the project's public README and repository metadata.