Add SSE transport to youtube-intelligence-mcp server

[ry-ops]

Overview

The youtube-intelligence-mcp server needs SSE (Server-Sent Events) implementation to enable remote access via HTTPS.

Current State

• Location: Needs verification (likely in services/ or related to youtube-channel-intelligence)
• Status: Pod running (1/1 Ready) as youtube-channel-intelligence-5994b75df5-tqkt6
• Port: 8081
• Ingress: Already configured at https://youtube-intelligence-mcp.ry-ops.dev/sse

Investigation Tasks

• Locate youtube-intelligence-mcp source code
• Determine current implementation (stdio vs HTTP)
• Check if SSE endpoint exists
• Review relationship with youtube-channel-intelligence deployment
• Test endpoint: curl https://youtube-intelligence-mcp.ry-ops.dev/sse

Implementation Requirements

• Add HTTP server with SSE support (if missing)
• Implement /sse endpoint for MCP protocol
• Bridge to YouTube intelligence functionality
• Add health check endpoint at /health
• Update Dockerfile to expose port 8081
• Ensure proper error handling

Current Pod Status

youtube-channel-intelligence-5994b75df5-tqkt6   1/1   Running   0   8m40s

Pod name suggests this may be the youtube-channel-intelligence service. Need to clarify:

• Is youtube-intelligence-mcp separate from youtube-channel-intelligence?
• Should they share the same SSE endpoint?
• Does ingress need updating to point to correct service?

Reference Implementations

Working MCP servers with SSE:

• proxmox-mcp-server
• sandfly-mcp-server
• unifi-mcp-server
• langflow-chat-mcp-server

Infrastructure Ready

✅ HTTPS ingress configured
✅ DNS resolving to youtube-intelligence-mcp.ry-ops.dev
✅ TLS certificate via Let's Encrypt
✅ Claude Desktop config includes this server
✅ Related pod running successfully

Testing Plan

1. Investigation Phase:

   • Locate source code
   • Review current implementation
   • Check logs: kubectl logs -n cortex youtube-channel-intelligence-5994b75df5-tqkt6
   • Verify service endpoints

2. Implementation Phase (if needed):

   • Add SSE wrapper to appropriate service
   • Build and push new image
   • Update deployment/ingress if needed
   • Deploy via ArgoCD

3. Verification Phase:

   • Test SSE endpoint with curl
   • Test with mcp-remote: npx -y mcp-remote https://youtube-intelligence-mcp.ry-ops.dev/sse
   • Verify in Claude Desktop app
   • Confirm YouTube intelligence features work

Questions to Resolve

1. Where is youtube-intelligence-mcp source code located?
2. Is it the same as youtube-channel-intelligence?
3. Does the ingress need to be updated to point to the correct service name?
4. What port should be exposed (8081 or different)?

Priority

High - Required for Claude Desktop YouTube intelligence features