menus-mcp
环球美食推荐 Plugin
一个 AI agent 插件:当用户表达就餐意向(「推荐晚餐」「不知道吃什么」「想吃点辣的」),基于 462 道菜的全球美食知识库做加权随机推荐。
零依赖:纯 Python 3 标准库
跨 agent:Claude Code(plugin)+ Gemini CLI(GEMINI.md)+ 任何 MCP 兼容 agent(规划中)
智能兜底:本地知识库 → 自动放宽 → 网络搜索 → LLM 常识,4 层兜底
去偏置 + 多样性:避免大菜系(如中餐)压制小菜系;避免推荐 3 道全是同一菜系
快速测试
python3 scripts/recommender.py --count 3 --keywords "辣,牛肉"返回 JSON:
{"dishes": [{"id": "...", "name": "麻婆豆腐", "price": 38, "cuisine": "川菜", ...}, ...], "exhausted": false, ...}完整参数:
python3 scripts/recommender.py --helpRelated MCP server: HowToCook-MCP Server
安装到不同 AI Agent
Claude Code
方式 A:自建 marketplace(推荐)
/plugin marketplace add spyyps/recommend-dish
/plugin install menus-recommender@spyyps-recommend-dish安装后在任意目录的 Claude Code 会话里说「推荐晚餐」「想吃点辣的」即可触发。
后续更新插件:/plugin update menus-recommender。
方式 B:本地 git clone
git clone https://github.com/spyyps/recommend-dish ~/.claude/plugins/menus-recommenderGemini CLI
git clone https://github.com/spyyps/recommend-dish
cd recommend-dish
gemini # 在本目录运行 Gemini CLI,会自动加载 GEMINI.mdCursor / Codex
待添加(参考 .cursor-plugin/ 与 .codex-plugin/ 的多 manifest 适配,第二阶段交付)。
MCP 兼容的任意 Agent
适用于 Claude Desktop / Cline / Cursor / Continue 等任何 MCP 客户端。
uvx 方式(推荐,无需 pip install):
{
"mcpServers": {
"menus": {
"command": "uvx",
"args": ["menus-mcp"]
}
}
}或者 pip install:
pip install menus-mcp{
"mcpServers": {
"menus": { "command": "menus-mcp" }
}
}触发示例
用户说 | 触发 | 解析为 |
「推荐晚餐」 | ✓ | 默认 3 道,无筛选 |
「想吃点辣的」 | ✓ |
|
「来 5 道便宜的海鲜」 | ✓ |
|
「推荐两道川菜」 | ✓ |
|
「换一批」(紧接上一次推荐) | ✓ |
|
「100 块以内的欧洲菜」 | ✓ |
|
算法核心
关键词硬过滤:用户传了关键词时,菜品必须至少命中一个(OR 语义)
多重命中加权:命中 N 个关键词的菜,权重 ×2^(N-1)
菜系反偏置:权重 ×1/√(该菜系总菜数),平衡中餐占 35% 的天然偏差
多样性贪心:同一菜系上限 ≤ ceil(count/3)
价格档松弛:候选不足自动去掉价格档重试一次
关键词白名单
只接受以下关键词,超出范围的(如「不辣」「清淡」)由 LLM 在调用前消化掉:
口味:麻辣 / 辣 / 甜 / 酸 / 咸 / 鲜
食材:海鲜 / 牛肉 / 羊肉 / 猪肉 / 鸡肉 / 素食
类型:面食 / 米饭类 / 汤 / 甜品 / 烧烤 / 火锅 / 咖喱 / 汉堡披萨 / 饮品
数据来源
全球菜系 462 道菜(10 个地区、41 个菜系、69 个子风味)
价格范围 ¥6 ~ ¥388(CNY)
完整数据见
menu.json与knowledge_base/各索引文件
发布到他人(仓库 owner 视角)
作为 Plugin 分享
git push到 GitHub(本仓库已就绪:spyyps/recommend-dish)仓库内
.claude-plugin/marketplace.json已配置好,列出所有可装插件在 README 贴出一行安装命令,分享仓库链接即可
作为 MCP 分享
仓库内 mcp-server/ 目录已实现 MCP server(Python,复用同一 recommender.py),暴露 recommend_dishes 与 list_keywords 两个工具。发布方式:
PyPI(推荐):
cd mcp-server pip install build twine python3 -m build twine upload dist/*用户
uvx menus-mcp或pip install menus-mcp即可。社区目录:发布完成后,提 PR 到
github.com/modelcontextprotocol/servers把 menus-mcp 加进列表。MCPB 单文件(Claude Desktop 专用):
npx @anthropic-ai/mcpb pack→ GitHub Release → 用户拖入 Claude Desktop 即装。
许可
MIT
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityFmaintenanceAn AI-powered server that helps users discover and book restaurants based on location, cuisine preferences, mood, and event type, with integration to Google Maps Places API for accurate recommendations.Last updated516MIT
- AlicenseBqualityDmaintenanceEnables AI assistants to recommend recipes, plan weekly meals, and solve the "what to eat today" problem by providing access to a comprehensive Chinese recipe database with smart filtering by categories, allergies, and dietary restrictions.Last updated58,253ISC
- Alicense-quality-maintenanceHelps users decide what to eat by randomly selecting meals from a database of 100+ options or a custom list using an interactive spinning wheel widget. It supports cuisine filtering and provides an engaging visual interface to resolve mealtime indecision.Last updated
- FlicenseAqualityDmaintenanceAn MCP server for restaurant discovery and booking across Resy and OpenTable via natural language. It integrates Google Places data with dietary preferences, visit history, and weather awareness to provide personalized dining recommendations and group reservation management.Last updated23
Related MCP Connectors
AI-native restaurant discovery: verified/menu-indexed/discovered tiers + signed allergy-safety data.
Semantic search across 50,000+ food recipes with hybrid retrieval and reranking.
Discover 356K+ restaurants in 20 countries: search, reviews, cultural context, time-based offers.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/spyyps/recommend-dish'
If you have feedback or need assistance with the MCP directory API, please join our Discord server