freestylefly

WeRead MCP Server is a tool that provides MCP (Model Context Protocol) services for WeRead, supporting the provision of book, note, and highlight data from WeRead to large language model (LLM) clients that support MCP, such as Cursor and Claude Desktop.

Features

• Fetch shelf information from WeRead
• Search for books on the shelf
• Retrieve notes and highlights from books
• Get popular reviews for books
• Organize notes and highlights by chapter
• Seamless integration with LLM clients that support the MCP protocol

Main Tools

1. get_bookshelf - Retrieves all books on the user's shelf

   • Returns basic book information including title, author, translator, and category, etc.

2. search_books - Searches for books on the user's shelf using keywords

   • Supports fuzzy and exact matching
   • Optional inclusion of detailed information
   • Can set the maximum number of results

3. get_book_notes_and_highlights - Retrieves all highlights and notes for a specified book

   • Supports organizing results by chapter
   • Allows filtering by highlight style
   • Returns structured data for easy LLM understanding

4. get_book_best_reviews - Retrieves popular reviews for a specified book

   • Supports setting the number of reviews to return
   • Supports pagination
   • Includes rating, likes, and reviewer information

Installation and Usage

Prerequisites

• Node.js 16.x or higher
• A WeRead account and valid Cookie

Installation Guide

For more details, see:

• Building a Personal Library in Cursor with WeRead MCP, Amazing!
• Weread MCP Server User Guide

Integration with Claude Desktop

There are several ways to integrate with Claude Desktop:

Method One: Using npx (Simplest, Recommended)

1. Open Claude Desktop

2. Go to Settings -> MCP Configuration

3. Add a tool using the following JSON configuration:
   json
   {
   "mcpServers": {
   "mcp-server-weread": {
   "command": "npx",
   "args": ["-y", "mcp-server-weread"],
   "env": {
   // Method 1: Using Cookie Cloud (Recommended)
   "CC_URL": "https://cc.chenge.ink", // Cookie Cloud URL
   "CC_ID": "Your ID", // Cookie Cloud ID
   "CC_PASSWORD": "Your Password" // Cookie Cloud Password

       // Or Method 2: Directly provide Cookie
       // "WEREAD_COOKIE": "Your WeRead Cookie"
     }
   }
   

   }
   }

Method Two: Use after global installation

1. Globally install the package:
   bash
   npm install -g mcp-server-weread

2. Use in Claude configuration:
   json
   {
   "mcpServers": {
   "mcp-server-weread": {
   "command": "mcp-server-weread",
   "env": {
   // Configure environment variables as above
   }
   }
   }
   }

Tip: Providing environment variables directly in the Claude configuration is more convenient and does not require setting up a .env file, which is recommended.

CookieCloud Configuration InstructionsTo address the issue of cookies frequently expiring and needing to be reacquired and updated in environment variables, this project supports the CookieCloud service for automatic synchronization and updating of cookies. CookieCloud is an open-source cross-browser cookie synchronization tool that supports self-hosted servers.

Configuration Steps:

Install the browser extension
Edge Store: CookieCloud for Edge
Chrome Web Store: CookieCloud for Chrome

Configure the CookieCloud Extension

Server Address: Use the default server https://cc.chenge.ink or enter your self-hosted server address.
Click "Auto-Generate Password"
In the Sync Domain Keywords, enter "weread"
Click "Save", then click "Manual Sync" to ensure the configuration takes effect.
[Optional] If you need the plugin to automatically maintain activity, enter https://weread.qq.com in the Keep Alive section, and the plugin will automatically refresh the cookie.

Configure CookieCloud variables in MCP Json:
CC_URL=Your CookieCloud server address (or use my default server address https://cc.chenge.ink)
CC_ID=Your CookieCloud user UUID
CC_PASSWORD=Your CookieCloud password
Note: After configuring CookieCloud, the system will prioritize using CookieCloud to obtain cookies. If it fails, it will use the value of the WEREAD_COOKIE environment variable.

Methods for Obtaining Cookies

Cookies for WeRead can be obtained through the following steps:

1. Log in to the WeRead web version using a Chrome browser.
2. Press F12 to open Developer Tools and switch to the Network tab.
3. Refresh the page and find the request for weread.qq.com in the request list.
4. In Headers, find the Cookie field and copy its entire content.
5. Add the copied content to the WEREAD_COOKIE variable in the MCP Json or .env file.

Usage Examples

Here are some examples of interactions with LLM:

Getting Bookshelf Information

User: Help me check my WeRead bookshelf.
LLM: I will fetch your WeRead bookshelf information.
[LLM calls the get_bookshelf tool]
I have retrieved 208 books from your WeRead bookshelf. Here are some of them:

1. "Incognito: The Secret Lives of the Brain" - David Eagleman
2. "Thinking, Fast and Slow" - Daniel Kahneman
3. "Sapiens: A Brief History of Humankind" - Yuval Noah Harari
   ...

Searching for Specific Books

User: Help me find books about psychology.
LLM: I will search for books related to psychology on your bookshelf.
[LLM calls the search_books tool with parameter keyword="psychology"]
On your bookshelf, I found 3 books related to psychology:

1. "Thinking, Fast and Slow" - Daniel Kahneman (Progress: 75%)
2. "The Crowd: A Study of the Popular Mind" - Gustave Le Bon (Progress: 50%)
3. "Psychology and Life" - Richard J. Gerrig (Progress: 30%)
   Which book would you like more information about?

Getting Notes and Highlights

User: Help me organize the notes for the book "Thinking, Fast and Slow".
LLM: I will fetch the notes and highlights for "Thinking, Fast and Slow".
[LLM calls the search_books tool to find the book ID]
[LLM calls the get_book_notes_and_highlights tool]
Here are your notes and highlights for "Thinking, Fast and Slow" (organized by chapter):

Chapter 1: System 1 and System 2

• Highlight: "System 1 operates automatically and quickly, with little or no effort and no sense of voluntary control."
• Your Note: "System 1 is responsible for intuition and quick judgments, while System 2 handles slow, rational thinking."

Chapter 3: Anchoring

• Highlight: "Anchoring occurs when people consider a particular value for an unknown quantity before estimating that quantity."
  ...

Getting Popular Book Reviews

User: What are the popular reviews for the book "Principles"?
LLM: I will find the popular reviews for "Principles".
[LLM calls the search_books tool to find the book ID]
[LLM calls the get_book_best_reviews tool]
Here are the popular reviews for "Principles":

1. Reviewer "Thinker" (Rating: 5 stars):
   "This is a book about how to think. Dalio summarizes his years of experience into actionable principles to help us make better decisions. I particularly like his view on 'Pain + Reflection = Progress', which is very practical."
   👍 182 | 💬 23

2. Reviewer "Investment Learner" (Rating: 4 stars):
   "The essence of the founder of Bridgewater Associates' thoughts, worth reading multiple times. The work principles section is especially helpful for managers. It's recommended to read the life principles first, followed by the work principles."
   👍 94 | 💬 12
   ...## Nine、Friendly Links

• mcp-server-weread : 🚀 A tool that provides MCP (Model Context Protocol) service for WeChat Reading, supporting the provision of books, notes, and highlighted data from WeChat Reading to large language model clients that support MCP, such as Claude Desktop.
• CodeCanvas : 📚 This code repository is a technical summary of the author Cang He's years of experience in front-line internet Java development. It aims to provide a clear and detailed learning tutorial, with a focus on writing core Java content. 💪🏻
• PmHub : 🔥 PmHub is an intelligent project management system based on SpringCloud & LLM microservices. This project is designed to help everyone quickly master the architectural design and development process of microservice/distributed projects. If you want to secure a satisfactory offer during campus or social recruitment, PmHub will be a very nice choice.

Ten、Acknowledgements

This project is forked from ChenyqThu's mcp-server-weread project, with some minor modifications, and it suddenly became popular. I also encourage everyone to star the original project on the author's GitHub. Additionally, I will synchronize any new updates from ChenyqThu, and the project will continue to update its usage tutorials. Thank you all for your support.

Eleven、Star Trend Chart

Twelve、Official Account

Search for Cang He on WeChat or scan the QR code below to follow Cang He's original official account. Reply with AI to discuss AI with over 5000 friends and learn about MCP together.

Thirteen、License

MIT License (MIT)
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

Copyright (c) 2025-2026 mcp-server-weread