🌐 Context Forge MCP Gateway - Use Cases & Capabilities

Federated MCP Gateway with Public Access via Cloudflare Tunnel

Public URL: http://mcp.deploystack.run

🎯 What You Now Have: Context Forge MCP Gateway Use Cases

Primary Use Case: AI-Powered Media & Download Management

You've deployed a production-ready federated MCP gateway that gives AI assistants like Claude, ChatGPT, and other MCP clients unified access to your entire media infrastructure through a single authenticated endpoint.

🚀 Practical Applications

1. AI Assistant Integration (Primary Use)

What It Enables:

Ask Claude: "Show me what's downloading in qBittorrent"
Ask ChatGPT: "Search for a specific TV series in Prowlarr"
Ask any MCP client: "What movies were recently added to Plex?"

How It Works:

// Add to Claude Code's .mcp.json
{
  "mcpServers": {
    "mcp-gateway": {
      "url": "http://mcp.deploystack.run",
      "transport": {
        "type": "http",
        "headers": {
          "Authorization": "Bearer YOUR_JWT_TOKEN"
        }
      }
    }
  }
}

Now Claude can directly control your 93 tools across all services without you writing any code.

2. Virtual MCP Servers (Custom Tool Combinations)

What You Can Build in the Web UI:

Example 1: "Media Discovery Server"

Combine: Prowlarr search tools + Radarr/Sonarr lookup + Plex library tools
Result: A custom MCP endpoint that helps find and add new content
Use Case: "Find and add 'Inception' to my media library"

Example 2: "Download Manager"

Combine: qBittorrent tools + SABnzbd tools + Prowlarr search
Result: Unified download control across torrent and Usenet
Use Case: "What's downloading and how much is left?"

Example 3: "Plex Admin Server"

Combine: Plex server info + user management + session control
Result: Server administration without opening Plex
Use Case: "Who's streaming right now and what are they watching?"

3. Team-Based Access Control

What It Enables:

Create separate teams with different tool access levels
Family members: Read-only Plex access
Power users: Full download management
Admins: Everything

Example Teams You Could Create:

Team	Access	Tools
Family	Plex library browsing only	10 Plex read tools
Downloads	Torrent/Usenet management	28 download tools
Admins	Full control	All 93 tools

4. API Access for Custom Applications

What You Can Build:

# Python script using your gateway
import requests

# Get JWT token
token = requests.post('http://mcp.deploystack.run/auth/login',
    json={'email': '[email protected]', 'password': 'pass'}).json()['access_token']

# Call any MCP tool
response = requests.post('http://mcp.deploystack.run/tools/plex-server-get-info',
    headers={'Authorization': f'Bearer {token}'})

print(f"Plex server has {response.json()['libraryCount']} libraries")

Use Cases:

Build mobile apps that control your media server
Create automation scripts (e.g., auto-pause torrents when bandwidth needed)
Integrate with home automation (Home Assistant, Node-RED)
Build Discord/Telegram bots for media requests

5. Agent-to-Agent (A2A) Communication

What It Enables:

AI agents can discover and communicate with each other
Build multi-agent workflows
Chain operations across services

Example A2A Workflow:

Search Agent finds media in Prowlarr
Download Agent adds to qBittorrent
Notification Agent monitors download
Plex Agent refreshes library when complete

📊 Your Current Deployment Stats

Category	Count	Examples
Total Tools	93	Cross-service operations
Plex Tools	51	Library, users, sessions, media
qBittorrent Tools	16	Torrent management
SABnzbd Tools	12	Usenet downloads
Prowlarr Tools	11	Indexer search
Radarr/Sonarr	3	Media lookup
Resources	2	Series & movies collections

🔥 Real-World Scenarios

Scenario 1: Media Request via Chat

You: "Hey Claude, find 'The Matrix' and add it to my Plex server"

Claude uses your gateway to:
1. Search Prowlarr for "The Matrix"
2. Add best result to qBittorrent
3. Monitor download progress
4. Notify when added to Plex

Scenario 2: Download Management

You: "What's downloading and how's my bandwidth?"

Claude checks:
- qBittorrent: Active torrents, speeds, ETA
- SABnzbd: Usenet queue status
- Returns unified status across both

Scenario 3: Plex Administration

You: "Who's streaming right now?"

Claude queries:
- Active Plex sessions
- What they're watching
- Bandwidth usage per stream
- Device information

🌐 Integration Options

Option 1: Claude Code (Desktop/Web)

Status: ✅ Ready to configure Config: Add gateway URL + JWT token to .mcp.json Result: Claude can use all 93 tools in conversation

Option 2: ChatGPT Actions

Status: ✅ Ready to import Steps: Import http://mcp.deploystack.run/openapi.json Result: ChatGPT can call your tools as custom actions

Option 3: Custom MCP Clients

Status: ✅ Ready for development Protocol: Standard MCP over HTTP Result: Build your own applications using the gateway

🏗️ Architecture

Internet
  ↓
Cloudflare Edge (DDoS Protection)
  ↓ QUIC/HTTP3 Tunnel
Proxmox Host (192.168.1.123)
  ↓ Nginx Reverse Proxy
LXC Container 106 (192.168.1.196:3000)
  ↓ MCP Gateway (FastAPI)
  ↓
Backend MCP Servers (SSE Protocol):
  ├─ Plex (localhost:3002/sse) - 51 tools
  ├─ qBittorrent (localhost:3003/sse) - 16 tools
  ├─ SABnzbd (localhost:3004/sse) - 12 tools
  ├─ Prowlarr (localhost:3005/sse) - 11 tools
  └─ Radarr/Sonarr (localhost:3006/sse) - 3 tools

🔐 Security Features

Active Protections

✅ JWT authentication (7-day expiration)
✅ Argon2 password hashing
✅ Bearer token authorization
✅ Cloudflare DDoS protection
✅ Encrypted tunnel (QUIC/HTTP3)
✅ Private backend network (LXC)

🎯 Bottom Line: What You Achieved

You now have a production-grade AI infrastructure hub that:

✅ Federates 5 separate services into one endpoint
✅ Exposes 93 tools to any MCP-compatible AI assistant
✅ Provides secure authentication (JWT + team-based access)
✅ Enables public internet access via Cloudflare Tunnel
✅ Allows custom virtual server creation for specific workflows
✅ Supports both AI assistants and custom applications

Practical Impact: Instead of manually checking Plex, qBittorrent, SABnzbd, and Prowlarr separately, you can now ask Claude (or any AI assistant) to manage your entire media infrastructure through natural conversation.

🔜 What You Could Do Next

Configure Claude Code: Add gateway to .mcp.json and start using tools in chat
Create Virtual Servers: Build custom tool combinations for specific use cases
Set Up Teams: Create team-based access for different users/applications
Build Custom Apps: Use the REST API to create mobile apps or automation scripts
Change Password: Update default password for security

📞 Quick Links

Resource	URL
Health Check	http://mcp.deploystack.run/health
API Documentation	http://mcp.deploystack.run/docs
OpenAPI Schema	http://mcp.deploystack.run/openapi.json
Login Endpoint	http://mcp.deploystack.run/auth/login
Tools Catalog	http://mcp.deploystack.run/tools

🚀 Quick Start

1. Get Access Token

curl -X POST http://mcp.deploystack.run/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email":"[email protected]","password":"Admin123!mcp"}' \
  | jq -r '.access_token'

2. Configure Claude Code

Add to your .mcp.json:

{
  "mcpServers": {
    "mcp-gateway": {
      "url": "http://mcp.deploystack.run",
      "transport": {
        "type": "http",
        "headers": {
          "Authorization": "Bearer YOUR_TOKEN_HERE"
        }
      }
    }
  }
}

3. Test It

Ask Claude: "List all MCP tools from mcp-gateway"

Expected: 93 tools listed! 🎉

Your gateway is fully operational and ready to use! 🎉

Status: 🟢 FULLY OPERATIONAL Last Verified: 2025-10-13 13:45 UTC

jmanhype/MCP_GATEWAY_USE_CASES.md

🌐 Context Forge MCP Gateway - Use Cases & Capabilities

🎯 What You Now Have: Context Forge MCP Gateway Use Cases

Primary Use Case: AI-Powered Media & Download Management

🚀 Practical Applications

1. AI Assistant Integration (Primary Use)

2. Virtual MCP Servers (Custom Tool Combinations)

3. Team-Based Access Control

4. API Access for Custom Applications

5. Agent-to-Agent (A2A) Communication

📊 Your Current Deployment Stats

🔥 Real-World Scenarios

Scenario 1: Media Request via Chat

Scenario 2: Download Management

Scenario 3: Plex Administration

🌐 Integration Options

Option 1: Claude Code (Desktop/Web)

Option 2: ChatGPT Actions

Option 3: Custom MCP Clients

🏗️ Architecture

🔐 Security Features

Active Protections

🎯 Bottom Line: What You Achieved

🔜 What You Could Do Next

📞 Quick Links

🚀 Quick Start

1. Get Access Token

2. Configure Claude Code

3. Test It