/mcp

Model Context Protocol server for AI agent integration

$ what is mcp

Model Context Protocol is a standard for connecting AI models to data sources and tools. Instead of calling REST APIs, your agent connects to an MCP server and accesses resources natively through its tool-calling interface.

MFM exposes the same catalog data available through the REST API, packaged as MCP resources and tools. Your agent can list services, fetch manifests, and run discovery queries — all through MCP.

$ available tools

mfm_list_services(domain?: string, skip?: int, limit?: int)

List all indexed service summaries. Returns service IDs, names, domains, and trust status. Supports optional domain and pagination filters.

mfm_get_service(service_id: string)

Fetch the full structured manifest for a service by ID. Includes capabilities, pricing, trust data, and integration details.

mfm_discover({ required_capabilities?, required_domains?, business_model?, trust_filter?, prefer?, exclude_services? })

Query the catalog with structured filters. Returns ranked matches with match strength, detail, and full query transparency.

mfm_list_taxonomy(domain?: string, group?: string)

List the controlled vocabulary used in MFM manifests. Discover valid terms for capabilities, frameworks, domains, etc.

$ connect

# stdio transport (local agents)
npx @mfm/mcp-server

# SSE transport (remote agents)
# Connect to: https://mcp.mfm.dev/sse

# Claude Desktop config (~/.claude/mcp_servers.json)
{
  "mfm": {
    "command": "npx",
    "args": ["@mfm/mcp-server"]
  }
}