AI-powered Zotero research assistant — a standalone MCP server with 36 tools for literature search, bilingual retrieval, reading analysis, citation management, and review writing.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"zra-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# Zotero Research Assistant [](https://www.python.org/) [](./LICENSE) [](https://modelcontextprotocol.io/) **[English](./README.md)** | **[中文](./README_zh.md)** --- > **Turn your Zotero library into an AI-searchable knowledge base.** > > A production-grade RAG pipeline — from PDF chunking to bilingual semantic retrieval — that runs entirely on your machine. Find papers by meaning, not just keywords. Works with any MCP-compatible AI client. --- ## Table of Contents - [RAG Pipeline](#rag-pipeline) — the core - [Quick Start](#quick-start) - [Client Setup](#client-setup) - [MCP Tools (36)](#mcp-tools-36) - [Configuration](#configuration) - [Tables & Figures](#tables--figures) - [Other Features](#other-features) - [Updating](#updating) - [Troubleshooting](#troubleshooting) - [Architecture](#architecture) - [Acknowledgments](#acknowledgments) - [License](#license) --- ## RAG Pipeline The RAG pipeline is the heart of this project. Every design decision — from chunking strategy to embedding backend to diversity reranking — is optimized for one goal: **maximize retrieval precision for academic papers on consumer hardware.** ### Pipeline Overview ``` Your Zotero Library │ ▼ ┌──────────────────────────────────────────────────────┐ │ 1. PDF EXTRACTION (PyMuPDF) │ │ Page-by-page text extraction, parallel processing │ ├──────────────────────────────────────────────────────┤ │ 2. TEXT CLEANING (52 regex rules) │ │ Strips journal boilerplate: article-info blocks, │ │ CLC numbers, funding footers, page numbers, DOIs │ │ EN journals (9 rules) · CN journals (24) · Univ. (19) │ │ Avg 10.6% line removal (CN 19.3%, EN 7.2%) │ ├──────────────────────────────────────────────────────┤ │ 3. SEMANTIC CHUNKING …
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.