explaingit

muisedestiny/zotero-gpt

7,128TypeScriptAudience · researcherComplexity · 2/5Setup · moderate

TLDR

Zotero-GPT is a plugin that connects your Zotero research library to OpenAI's AI models so you can summarize papers, ask questions about PDFs, and search your library with AI without leaving Zotero.

Mindmap

mindmap
  root((repo))
    What it does
      Zotero AI plugin
      Chat with papers
      Library search
    Command tags
      Built-in prompts
      Custom JavaScript
      Trigger word activation
    AI models
      GPT-3.5-turbo
      GPT-4 support
      Custom API endpoint
    Interface
      Floating response window
      Markdown rendering
      LaTeX math support
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Ask an AI to summarize any paper you have open in Zotero without switching to another app or copy-pasting text.

USE CASE 2

Search your entire Zotero library with AI by selecting a passage and running a built-in library search tag.

USE CASE 3

Write custom command tags in JavaScript to automate recurring tasks like extracting methods sections from papers.

USE CASE 4

Use GPT-4 to answer questions about a PDF you are reading in Zotero, with LaTeX math rendered in the floating response window.

Tech stack

TypeScriptJavaScript

Getting it running

Difficulty · moderate Time to first run · 30min

Requires your own OpenAI API key, which incurs usage costs per request.

The explanation does not specify the license terms.

In plain English

Zotero-GPT is a plugin for Zotero, the free reference management application used by researchers and academics to collect, organize, and cite papers. The plugin connects Zotero to OpenAI's language models, letting you ask questions about papers in your library and get AI-generated answers without leaving the application. The core interaction model is built around what the project calls command tags. These are saved prompts that you assign a name and a trigger word. Once set up, you can click a tag or type its trigger word to instantly run that prompt against whatever you currently have selected, whether that is the full text of an open PDF, a selected passage, or the abstract of a paper in your library. Built-in tags cover common research tasks: summarizing a paper, answering questions about a PDF, and searching your library based on selected text. You can also write your own tags using JavaScript code that calls into Zotero's own API, which allows for more complex interactions. The plugin works with both GPT-3.5-turbo and GPT-4. You provide your own OpenAI API key in the plugin settings, where you can also adjust parameters like the model name, API endpoint, and response temperature. The floating window that displays responses supports real-time markdown rendering including LaTeX math notation, can be moved anywhere on screen, and allows zooming the text in or out. Installation follows the standard Zotero plugin process: download the .xpi file from the releases page and load it through Zotero's add-ons menu. The plugin is compatible with both Zotero 6 and Zotero 7. A community discussion thread on the repository is used to share command tag recipes contributed by other users.

Copy-paste prompts

Prompt 1
I installed Zotero-GPT and want to create a command tag that extracts the methodology section from any paper I open. Write the JavaScript code for that tag.
Prompt 2
How do I configure Zotero-GPT to use GPT-4 instead of GPT-3.5-turbo, and where do I enter my OpenAI API key?
Prompt 3
Write a Zotero-GPT command tag that finds all papers in my library related to the text I currently have selected.
Prompt 4
Walk me through installing Zotero-GPT on Zotero 7 by downloading the .xpi file and loading it through the add-ons menu.
Open on GitHub → Explain another repo

← muisedestiny on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.