LinkVault MCP服务器
一个用于扩展Amazon Q CLI的工具，具有AI辅助书签管理功能。
概述
LinkVault是一个MCP（模型上下文协议）服务器，与Amazon Q CLI集成，通过自然语言提供智能书签管理。它允许您在终端中使用对话命令访问、组织和搜索书签。
主要特点
- Amazon Q CLI集成：在终端中使用自然语言管理书签
- Chrome多配置文件支持：从所有Chrome配置文件访问书签
- 内容提取：自动分析网页内容以进行更好的分类
- 智慧型组织：人工智能辅助分类和标记
- 交叉平台的：适用于macOS、Windows和Linux
安装
- 请确保您已安装Python 3.10+
- 安装所需的软件包：
pip install -r requirements.txt- 使脚本可执行：
chmod +x main.py src/url_manager.py src/server.pyAmazon Q CLI设置
要将LinkVault与Amazon Q CLI一起使用，您需要将其注册为MCP服务器：
- 编辑MCP配置文件：
nano ~/.aws/amazonq/mcp.json- 添加书签管理器配置：
"bookmark_manager": {
"command": "uv",
"args": ["--directory", "/path/to/linkvault-mcp-server", "run", "src/server.py"],
"env": {},
"disabled": false,
"autoApprove": ["get_url_data", "store_url", "search_bookmarks", "list_categories", "list_bookmarks_by_category", "delete_bookmark", "list_chrome_bookmarks", "import_chrome_bookmark"]
}- 替换
/path/to/linkvault-mcp-server带有安装目录的实际路径。
- 保存文件并重新启动Amazon Q CLI（如果它已在运行）。
MCP架构
LinkVault使用模型上下文协议来扩展具有书签管理功能的Amazon Q CLI：
graph TD
User[User] -->|Natural Language| AQ[Amazon Q CLI]
AQ -->|MCP Protocol| MCP[LinkVault MCP Server]
MCP -->|SQLite Storage| SQL[(SQLite Database)]
MCP -->|Extract Content| Web[Web Pages]
MCP -->|Read Bookmarks| Chrome[Chrome Profiles]
subgraph LinkVault
MCP
endMCP工作流程
sequenceDiagram
participant User
participant AQ as Amazon Q CLI
participant MCP as LinkVault MCP Server
participant Web as Web Pages
participant DB as SQLite Database
participant Chrome as Chrome Bookmarks
User->>AQ: "Save this article: https://example.com"
AQ->>MCP: get_url_data(url)
MCP->>Web: Fetch content
Web-->>MCP: HTML content
MCP-->>AQ: URL metadata
AQ->>User: "I found this article about X. Save it?"
User->>AQ: "Yes, save it"
AQ->>MCP: store_url(metadata)
MCP->>DB: Store bookmark
DB-->>MCP: Success
MCP-->>AQ: Success response
AQ->>User: "Bookmark saved!"
User->>AQ: "Show my Chrome bookmarks"
AQ->>MCP: list_chrome_bookmarks()
MCP->>Chrome: Read bookmarks
Chrome-->>MCP: Bookmark data
MCP-->>AQ: Chrome bookmarks
AQ->>User: "Here are your Chrome bookmarks..."浏览器集成
LinkVault可以从系统上的所有Chrome配置文件访问书签：
flowchart TD
start[Start] --> findProfiles[Find Chrome Profiles]
findProfiles --> profileCheck{Profiles Found?}
profileCheck -->|No| returnEmpty[Return Empty List]
profileCheck -->|Yes| loopProfiles[Process Each Profile]
loopProfiles --> parseBookmarks[Parse Bookmark File]
parseBookmarks --> extractData[Extract Bookmark Data]
extractData --> addProfileInfo[Add Profile Information]
addProfileInfo --> combineResults[Combine Results]
combineResults --> filterCheck{Filter by Folder?}
filterCheck -->|Yes| applyFilter[Apply Folder Filter]
filterCheck -->|No| returnAll[Return All Bookmarks]
applyFilter --> returnFiltered[Return Filtered Bookmarks]
returnEmpty --> endNode[End]
returnAll --> endNode
returnFiltered --> endNode与Amazon Q CLI的交互示例
$ q chat
Welcome to Amazon Q CLI!
> Save this article: https://docs.aws.amazon.com/lambda/latest/dg/lambda-urls.html
I'll analyze that article for you. One moment...
I've analyzed the article and it appears to be about AWS Lambda Function URLs.
Would you like me to save it with the following details?
- Category: Serverless
- Tags: lambda, api, function-urls, http-endpoints
- Description: Documentation on how to create and configure function URLs to provide dedicated HTTP(S) endpoints for Lambda functions.
- Importance: 4 (High)
> Yes, that looks good
Great! I've saved the article to your bookmarks.
> Show me my Chrome bookmarks about AWS
I'll check your Chrome bookmarks for AWS-related content.
Found 12 bookmarks related to AWS across your Chrome profiles:
1. AWS Management Console
URL: https://console.aws.amazon.com/
Profile: Work/Bookmarks Bar
2. Amazon EC2 Documentation
URL: https://docs.aws.amazon.com/ec2/
Profile: Work/AWS Documentation
[...more results...]
Would you like me to import any of these into your LinkVault collection?
> Import the first one into my AWS category
I've imported "AWS Management Console" into your AWS category with tags: aws, console, management.MCP工具参考
获取日期
def get_url_data(url: str) -> Dict[str, Any]:
"""
Fetch and extract data from a URL.
Args:
url: The URL to fetch and analyze
Returns:
A dictionary containing extracted data from the URL
"""store_url
def store_url(url: str, title: str, category: str,
tags: List[str], description: str,
importance: int, notes: str = None) -> Dict[str, Any]:
"""
Store a URL with AI-generated metadata.
Args:
url: The URL to store
title: The title of the webpage
category: The category to assign
tags: List of tags to associate with the URL
description: A brief description of the content
importance: Importance rating (1-5)
notes: Optional additional notes or comments about the URL
Returns:
A dictionary indicating success or failure
"""搜索_书签
def search_bookmarks(query: str) -> Dict[str, Any]:
"""
Search for bookmarks by query.
Args:
query: The search query
Returns:
A dictionary containing search results
"""列表_类别
def list_categories() -> Dict[str, Any]:
"""
List all bookmark categories.
Returns:
A dictionary containing all categories and their counts
"""list_bookmarks_by_category
def list_bookmarks_by_category(category: str) -> Dict[str, Any]:
"""
List all bookmarks in a specific category.
Args:
category: The category name
Returns:
A dictionary containing bookmarks in the specified category
"""删除书签
def delete_bookmark(url: str, category: str = None) -> Dict[str, Any]:
"""
Delete a bookmark from the database.
Args:
url: The URL of the bookmark to delete
category: Optional category to specify which bookmark to delete if the same URL exists in multiple categories
Returns:
A dictionary indicating success or failure and details about the deleted bookmark
"""list_chrome_bookmarks
def list_chrome_bookmarks(folder: str = None) -> Dict[str, Any]:
"""
List Chrome bookmarks, optionally filtered by folder.
Args:
folder: Optional folder path to filter bookmarks
Returns:
A dictionary containing Chrome bookmarks
"""进口_彩色_书签
def import_chrome_bookmark(url: str, title: str, category: str,
tags: List[str], description: str = "",
importance: int = 3, notes: str = None) -> Dict[str, Any]:
"""
Import a Chrome bookmark into the database.
Args:
url: The URL of the Chrome bookmark to import
title: The title of the bookmark
category: The category to assign
tags: List of tags to associate with the URL
description: A brief description of the content
importance: Importance rating (1-5)
notes: Optional additional notes or comments about the URL
Returns:
A dictionary indicating success or failure
"""数据存储
LinkVault将书签存储在SQLite数据库中 ~/Documents/github/linkvault-mcp-server/data/bookmarks.db
发展
需求
- Python 3.10+
- pyproject.toml中的依赖关系：
- 美容组4 - 快速 - fastmcp - 主控程序 - 请求：
项目结构
linkvault-mcp-server/
├── src/ # Source code directory
│ ├── __init__.py # Package initialization
│ ├── server.py # MCP server implementation
│ ├── url_manager.py # CLI implementation
│ └── utils/ # Utility functions
│ ├── __init__.py
│ └── browser_integration.py # Browser integration utilities
├── data/ # Database storage
├── main.py # Main entry point
├── README.md # This file
├── CHANGELOG.md # Version history
└── pyproject.toml # Project dependencies附加功能
命令行接口命令模式
除了MCP与Amazon Q CLI的集成外，LinkVault还提供了一个独立的CLI界面，用于直接书签管理。
./main.py --mode cli或者直接：
./src/url_manager.py [options]CLI命令
- 添加URL:
add [-t tags] [-n notes] [--title custom_title] - 列表类别:
categories - 按类别列出URL:
list - 搜索URL:
search - 列表标记:
tags - 列出带有标签的URL:
tag - 删除URL:
delete [-c category] - 重命名类别: `rename
`
- 删除类别:
delcat - 列出Chrome书签:
chrome [-f folder] - 导入Chrome书签:
import [-t tags] [--title title]
CLI数据存储
CLI模式使用JSON文件存储在 ~/Documents/github/linkvault-mcp-server/data/url_database.json
许可证
MIT许可证