Skip to content
MCP server

Menus Recommender

MCP server that suggests diverse dishes from a global food knowledge base for AI clients

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI, API
Runs
Locally
Version
1.0.1
Listing updated
today

Similar MCP servers

About Menus Recommender

Menus Recommender is an open-source MCP server for food and drink. It runs locally on your machine and is started by your AI client.

The README says this server ships a weighted random dish recommender built on a knowledge base of roughly 460 dishes across 41 cuisines, so an AI assistant can propose meal ideas instead of generating them from scratch. Callers can narrow results by keywords such as flavour, ingredient or dish type, by price tier or price ceiling in CNY, by cuisine or broad region, and can pass previously shown dish IDs to avoid repeats when asking for a fresh batch.

The main tool, recommend_dishes, returns structured JSON with dish names in Chinese and English, price, cuisine, region and a short description, plus a flag indicating whether the candidate pool was exhausted. The README describes a fallback chain for that case: relaxing the price filter, searching the web, or answering from general knowledge, with the calling agent responsible for that logic.

It is aimed at developers and users of MCP-compatible clients such as Claude Desktop, Cline, Cursor and Continue who want dining suggestions. It is installable from PyPI via uvx or pip, or can be run from source.

Install

Run in a terminal

claude mcp add menus-recommender -- uvx menus-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "menus-recommender": {
      "command": "uvx",
      "args": [
        "menus-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "menus-recommender": {
      "command": "uvx",
      "args": [
        "menus-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "menus-recommender": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "menus-mcp"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "menus-recommender": {
      "command": "uvx",
      "args": [
        "menus-mcp"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "menus-recommender": {
      "command": "uvx",
      "args": [
        "menus-mcp"
      ]
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Tools this server provides (2)

Taken from the project documentation; the running server may differ.

recommend_dishes
Returns weighted random dish recommendations with optional keyword, price, cuisine, region and exclusion filters.
list_keywords
Returns the supported keyword whitelist.

What it does

Weighted random recommendations
Picks dishes at random with weights that offset large cuisines so smaller ones still appear.
Keyword and price filters
Filters results by a fixed keyword list covering taste, ingredient and dish type, plus price tiers and a maximum price.
Cuisine and region selection
Narrows suggestions to a named cuisine or a broad geographic region.
Deduplicated refresh
Accepts excluded dish IDs so repeated requests return a different set of dishes.
Structured JSON output
Returns dish IDs, names in Chinese and English, prices, cuisines, regions and descriptions as JSON.
Exhaustion signal
Reports when the candidate pool is too small so the calling agent can fall back to other sources.

Good for

  • Suggesting dinner options within a set budget
  • Finding dishes from a specific cuisine or region for a meal plan
  • Generating a fresh list of dish ideas without repeating earlier picks
  • Giving an AI assistant a local knowledge base to answer what to eat questions

Pricing

Open source

Free to install from PyPI or source; no paid plan mentioned.

Menus Recommender FAQ

How do I install the server?

The README lists three methods: run it with uvx using the PyPI package menus-mcp, install it with pip and point the client at the menus-mcp command, or run the source file from the GitHub repository.

Does it need any API keys or environment variables?

The registry entry lists no environment variables for the PyPI package.

What happens when a keyword is passed?

The README says keywords act as hard filters with OR matching, so a dish must match at least one supplied keyword; it advises omitting a keyword rather than passing a negated one.

Which MCP tools are exposed?

The README documents recommend_dishes and list_keywords.

Is Menus Recommender free?

Yes. Menus Recommender is open source, so you can use and self-host it for free.

How do I add Menus Recommender to Claude Code?

Run this command in your terminal: claude mcp add menus-recommender -- uvx menus-mcp

What are the best alternatives to Menus Recommender?

The closest alternatives to Menus Recommender are Spoonacular, MCP Ptitchef, MCP Recipes, ask.recipes and Frigolog HACCP — French Food Safety Compliance.

Where this comes from

  • Description: written from the project's README and its MCP Registry entry.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers