LabStore
← All labs

Plugged.in Proxy

by docker ✓ verified ✓ Packaged by Docker

A unified MCP proxy that aggregates multiple MCP servers into one interface, enabling seamless tool discovery and management across all your AI interactions. Manage all your MCP servers from a single connection point with RAG capabilities and real-time notifications.

MCP server installable needs setup network: on runtime: sysbox

Project Management MCP Servers

Plugged.in Proxy MCP Server

A unified MCP proxy that aggregates multiple MCP servers into one interface, enabling seamless tool discovery and management across all your AI interactions. Manage all your MCP servers from a single connection point with RAG capabilities and real-time notifications.

What is an MCP Server?

MCP Info

Attribute Details
Docker Image mcp/pluggedin-mcp-proxy
Author VeriTeknik
Repository https://github.com/VeriTeknik/pluggedin-mcp-proxy

Image Building Info

Attribute Details
Dockerfile https://github.com/VeriTeknik/pluggedin-mcp-proxy/blob/ea42ab826a97b0f6bd49a582d024d69d4957c868/Dockerfile
Commit ea42ab826a97b0f6bd49a582d024d69d4957c868
Docker Image built by Docker Inc.
Docker Scout Health Score Docker Scout Health Score
Verify Signature COSIGN_REPOSITORY=mcp/signatures cosign verify mcp/pluggedin-mcp-proxy --key https://raw.githubusercontent.com/docker/keyring/refs/heads/main/public/mcp/latest.pub
Licence Apache License 2.0

Available Tools (6)

Tools provided by this Server Short Description
pluggedin_ask_knowledge_base Ask questions and get AI-generated answers from your knowledge base.
pluggedin_delete_notification Delete a notification from the Plugged.in system
pluggedin_discover_tools Triggers discovery of tools (and resources/templates) for configured MCP servers in the Pluggedin App.
pluggedin_list_notifications List notifications from the Plugged.in system with optional filters for unread only and result limit
pluggedin_mark_notification_done Mark a notification as done in the Plugged.in system
pluggedin_send_notification Send custom notifications through the Plugged.in system with optional email delivery.

Tools Details

Tool: pluggedin_ask_knowledge_base

Ask questions and get AI-generated answers from your knowledge base. Returns structured JSON with answer, document sources, and metadata.

Parameters Type Description
query string Your question or query to get AI-generated answers from the knowledge base.

Tool: pluggedin_delete_notification

Delete a notification from the Plugged.in system

Parameters Type Description
notificationId string The ID of the notification to delete

Tool: pluggedin_discover_tools

Triggers discovery of tools (and resources/templates) for configured MCP servers in the Pluggedin App.

Parameters Type Description
force_refresh boolean optional Set to true to bypass cache and force a fresh discovery. Defaults to false.
server_uuid string optional Optional UUID of a specific server to discover. If omitted, attempts to discover all.

Tool: pluggedin_list_notifications

List notifications from the Plugged.in system with optional filters for unread only and result limit

Parameters Type Description
limit integer optional Limit the number of notifications returned (1-100)
onlyUnread boolean optional Filter to show only unread notifications

Tool: pluggedin_mark_notification_done

Mark a notification as done in the Plugged.in system

Parameters Type Description
notificationId string The ID of the notification to mark as read

Tool: pluggedin_send_notification

Send custom notifications through the Plugged.in system with optional email delivery. You can provide a custom title or let the system use a localized default.

Parameters Type Description
message string The notification message content
sendEmail boolean optional Whether to also send the notification via email
severity string optional The severity level of the notification (defaults to INFO)
title string optional Optional notification title. If not provided, a localized default will be used. Consider generating a descriptive title based on the message content.

Use this MCP Server

{
  "mcpServers": {
    "pluggedin-mcp-proxy": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "PLUGGEDIN_API_BASE_URL",
        "-e",
        "PLUGGEDIN_API_KEY",
        "mcp/pluggedin-mcp-proxy"
      ],
      "env": {
        "PLUGGEDIN_API_BASE_URL": "https://plugged.in",
        "PLUGGEDIN_API_KEY": "YOUR_PLUGGEDIN_API_KEY"
      }
    }
  }
}

Why is it safer to run MCP Servers with Docker?