Bring your agent. Learn together.

What is MCP?

The Model Context Protocol lets AI agents call external tools and read resources through a standardized interface. Instead of browsing a website, your agent connects directly and receives structured data.

Server endpoint

Transport: Streamable HTTP

Endpoint URL
https://liquidgold.cc/mcp

Claude Desktop

Add this to your claude_desktop_config.json:

claude_desktop_config.json
{
  "mcpServers": {
    "liquidgold": {
      "url": "https://liquidgold.cc/mcp"
    }
  }
}

Cursor

Add this to .cursor/mcp.json:

.cursor/mcp.json
{
  "mcpServers": {
    "liquidgold": {
      "url": "https://liquidgold.cc/mcp"
    }
  }
}

MCP App

This is a registered MCP server that AI agents can discover and connect to.

Server name

liquidgold

Version

2.0.0

Transport

Streamable HTTP

Info endpoint
GET /api/v1/mcp/info
https://liquidgold.cc/api/v1/mcp/info

Tools

  • list_capabilities

    List every available capability. Returns the full toolkit catalogue.

  • search_capabilities

    Filter the capability catalogue by words and tags.

  • get_capability_schemas

    Retrieve JSON schemas for one or more capabilities.

  • execute_capabilities

    Execute one or more capabilities by name.

Resources

  • liquidgold://channel/{handle}/profile

    Read a channel's public profile (title, description, subscriber count, avatar)

  • liquidgold://channel/{handle}/videos

    List a channel's public videos with pagination

  • liquidgold://video/{videoId}

    Read one video's detail including curated overlay

  • liquidgold://categories

    List all channel categories

  • liquidgold://platform-metrics

    Read platform-wide metrics (channel count, video count)