pluggedin-app

概述

plugged.in 是一个 AI 内容管理系统，它将短暂的 AI 交互转变为持久、可版本化和可搜索的组织知识。它充当连接各种 AI 客户端与您的知识库以及更广泛的 MCP 生态系统的中央枢纽。该平台包含一个嵌入式的 RAG 向量引擎，用于文档处理和语义搜索，多模型协作跟踪，以及与 1,500 多个服务器的通用 MCP 集成。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• create_documentcreate_document(title: string, content: string, source?: string)

  Create a new document with version tracking and model attribution

  试试:Create a document called 'Project Analysis' with the content we discussed试试:Save this research as a new document titled 'Market Findings'
• query_ragquery_rag(query: string, limit?: number)

  Query the RAG knowledge base with semantic search

  试试:Find documents related to our authentication system试试:Search for information about the latest API changes
• re_index_vectorsre_index_vectors(document_id?: string)

  Trigger re-indexing of document vectors for corrupted or missing embeddings

  试试:Re-index all document vectors试试:Fix the vectors for the 'API Documentation' document
• list_clipboard_entrieslist_clipboard_entries(visibility?: 'private' | 'workspace' | 'public')

  List all stored clipboard entries with optional filtering by visibility

  试试:Show me all workspace clipboard entries试试:List all public clipboard items
• get_clipboard_entryget_clipboard_entry(name: string)

  Retrieve a specific clipboard entry by name

  试试:Get the 'customer_context' clipboard entry试试:Show me the 'last_analysis' data
• set_clipboard_entryset_clipboard_entry(name: string, value: any, visibility?: 'private' | 'workspace' | 'public', ttl?: number)

  Store or update a clipboard entry with optional expiration and visibility settings

  试试:Save this analysis as 'current_analysis'试试:Store the config with 1-hour expiration
• pop_clipboard_entrypop_clipboard_entry(name: string)

  Remove and return the last entry from a named clipboard stack

  试试:Pop the last task from the 'todo' stack试试:Remove and return the latest analysis
• push_clipboard_entrypush_clipboard_entry(name: string, value: any)

  Add a new entry to a named clipboard stack

  试试:Push this task to the 'todo' stack试试:Add the new analysis to the history stack
• list_mcp_serverslist_mcp_servers(types?: string[], search?: string)

  Discover and list available MCP servers with advanced filtering options

  试试:Show all Python MCP servers试试:Find MCP servers related to data analysis
• test_mcp_servertest_mcp_server(server_id: string, tool_name: string, args?: any)

  Test a specific tool from an MCP server in the interactive playground

  试试:Test the 'query_postgres' tool from the analytics server试试:Try the 'generate_report' function
• get_mcp_toolsget_mcp_tools(server_id: string)

  Retrieve all tools available from a specific MCP server

  试试:Show all tools from the GitHub integration server试试:List available functions in the analytics server
• log_mcp_activitylog_mcp_activity(server_id: string, tool_name: string, request: any, response: any)

  Log detailed MCP interactions for debugging and analytics

  试试:Log this API call to the database server试试:Record the interaction with the file system tool

可对比工具

安装

git clone https://github.com/VeriTeknik/pluggedin-app.git
cd pluggedin-app
cp .env.example .env

docker compose up --build -d

{
  "mcpServers": {
    "pluggedin": {
      "command": "docker",
      "args": ["run", "--rm", "-i", "veriteknik/pluggedin:latest", "mcp"]
    }
  }
}

FAQ

plugged.in 支持哪些平台？

plugged.in 通过 Docker 支持 amd64 和 arm64 架构，具有自动平台检测功能。它与 Claude Desktop、Cline、LM Studio 和其他兼容 MCP 的客户端配合工作。

plugged.in 如何处理数据存储和安全？

该平台使用 PostgreSQL 18 和 pgvector 进行数据库存储，并使用嵌入式 zvec 进行向量搜索。所有敏感数据均使用 AES-256-GCM 加密，并具有按配置文件的加密密钥。它还支持 OAuth 2.1 身份验证。

pluggedin-app 对比