Nom MCP server

Nom runs a remote Model Context Protocol server, so AI assistants can look up fast food nutrition and Nom Scores as native tools — "what's the healthiest thing at Chipotle?" or "compare a Big Mac and a Whopper" — answered straight from Nom's data.

Server URL https://mcp.nom.now/mcp
Remote MCP over Streamable HTTP (SSE at /sse). No auth, no key. It serves the same data as the JSON API, always in sync.

Connect it

Claude (Desktop or claude.ai)

Settings → Connectors → Add custom connector, and paste the server URL https://mcp.nom.now/mcp.

Claude Code

claude mcp add --transport http nom https://mcp.nom.now/mcp

Cursor, Windsurf, and other clients

Add Nom to your client's MCP config:

{
  "mcpServers": {
    "nom": { "url": "https://mcp.nom.now/mcp" }
  }
}

For a client that only speaks stdio, bridge to the remote server with mcp-remote: npx -y mcp-remote https://mcp.nom.now/mcp.

Tools

search_menu_items

Search the catalog by name or category; scope to one chain.

get_menu_item

Full nutrition + Nom Score for one item id.

get_nom_score

The Nom Score breakdown — per-component ratings.

list_restaurants

Every chain Nom covers, with item counts.

get_restaurant_menu

One chain's full menu with grades.

healthiest_items

Top items by Nom Score, by chain and/or diet.

compare_items

Several items side by side, fully scored.

Prefer raw HTTP? Everything here is also a plain JSON API at https://nom.now/api/v1.