Skip to main content
Glama
Bar Assistant MCP Server

Bar Assistant MCP Server

by zhdenny

Bar Assistant MCP Server

A high-performance Model Context Protocol (MCP) server that provides intelligent cocktail search and recipe retrieval from Bar Assistant instances.

Features

  • ๐Ÿ” Smart Cocktail Search - Natural language queries with similarity matching

  • ๐Ÿ“– Complete Recipes - Detailed ingredients, instructions, and specifications

  • ๐ŸŽฏ Similarity Engine - Discover cocktails similar to your favorites

  • ๐Ÿงช Ingredient Analysis - Find cocktails by ingredients, flavors, and characteristics

  • โšก Batch Processing - Retrieve multiple recipes simultaneously (5-10x faster)

  • ๐Ÿ’พ Smart Caching - Intelligent caching with 70%+ hit rate

  • ๐Ÿ”ง Advanced Filtering - ABV ranges, glassware, preparation methods, and more

Related MCP server: Cocktail By Api Ninjas

Prerequisites

  • Docker and Docker Compose

  • Access to a Bar Assistant instance

  • Bar Assistant API token

Running with Docker

The only supported method for running the server is with Docker Compose.

  1. Create an .env file:

    Create a file named .env in the root of the project and add the following, replacing the placeholder values with your actual Bar Assistant configuration:

    BAR_ASSISTANT_URL=https://your-instance.com
    BAR_ASSISTANT_TOKEN=your-api-token
    BAR_ASSISTANT_BAR_ID=1
  2. Get Your Bar Assistant API Token:

    1. Log into your Bar Assistant instance

    2. Navigate to Settings โ†’ API

    3. Generate a new API token

    4. Copy the token and paste it into the BAR_ASSISTANT_TOKEN field in your .env file.

  3. Build and start the container:

    docker-compose up --build -d

    The server will be available at http://localhost:3001.

PourOver API Gateway

This server includes an integrated API Gateway to support the PourOver mobile application, providing natural language query streaming using Google Antigravity CLI (agy).

  • Endpoint: POST /query

  • Authentication: Secured with MCP_SSE_TOKEN passed via the x-api-key header (or standard Authorization/query parameters).

  • Format: Takes a JSON body {"query": "your query"} and streams response chunks back in text/plain.

Usage

Once configured, you can interact with the server through your MCP client. Here are some example queries:

  • "What cocktails can I make with gin and vermouth?"

  • "Show me the recipe for a Manhattan"

  • "Give me recommendations on cocktails like a Negroni"

  • "What ingredients do I need to buy to make these 5 cocktails?"

Available Tools

3 tools
get_ingredient_infoA

Get comprehensive information about cocktail ingredients and their usage.

Use Cases:

  • Ingredient research: "what is Aperol?", "tell me about gin"

  • Substitution guidance: finding alternatives for unavailable ingredients

  • Usage exploration: see how ingredients are used across different cocktails

  • Flavor profile understanding: learn about ingredient characteristics

Response Format: Returns detailed ingredient information including:

  • Ingredient description and characteristics

  • List of cocktails using this ingredient (with complete recipes)

  • Suggested substitutions with flavor impact notes

  • Common flavor profiles and tasting notes

  • Direct links to featured cocktails

Examples:

  • {ingredient_name: "Campari"} โ†’ Campari info + Negroni, Boulevardier recipes

  • {ingredient_name: "rye whiskey"} โ†’ Usage in Manhattan, Sazerac, etc.

  • {ingredient_name: "elderflower liqueur"} โ†’ Aviation, Paper Plane recipes

Inputs: ingredient_name (required). Returns: query, metadata, ingredient, description, +3 more

get_recipeA

๐Ÿธ Advanced recipe retrieval with powerful batch processing for multiple cocktails.

๐Ÿš€ BATCH PROCESSING SYSTEM:

  • High Performance: 5-10x faster than sequential requests

  • Parallel Processing: Simultaneous API calls with error isolation

  • Smart Caching: 70%+ cache hit rate for repeated requests

  • Flexible Input: Mix cocktail names and IDs in single request

  • Error Resilience: Individual failures don't break entire batch

๐Ÿ“‹ LLM Usage Patterns:

  • Single Recipe: When user asks for "how to make [cocktail]"

  • Recipe Comparison: When user wants to compare multiple cocktails

  • Menu Planning: Batch retrieve recipes for event planning

  • Variation Exploration: Get base recipe + similar cocktails

  • Research Mode: Efficient lookup of multiple specific recipes

๐ŸŽฏ Input Methods (Choose Based on Use Case):

  1. Single Recipe (Backwards Compatible):

    • cocktail_name: "Manhattan" โ†’ One complete recipe

    • cocktail_id: 123 โ†’ Recipe by database ID

  2. Batch by Names (Most Common):

    • cocktail_names: ["Negroni", "Manhattan", "Martini"] โ†’ Multiple complete recipes

  3. Batch by IDs (When Available):

    • cocktail_ids: [1, 2, 3] โ†’ Multiple recipes by database IDs

  4. Mixed Batch (Maximum Flexibility):

    • cocktail_names: ["Aviation"] + cocktail_ids: [123, 456] โ†’ Combined approach

  5. With Variations (Exploration):

    • Any above + include_variations: true โ†’ Base recipes + similar cocktails

๐Ÿ“Š Response Format: Structured output with complete recipe data:

  • Precise ingredient measurements (auto-converted to oz)

  • Step-by-step preparation instructions

  • Cocktail specifications (ABV, glassware, method, garnish)

  • Direct database links for each recipe

  • Performance metrics (timing, cache usage)

  • Similar recipes when requested

  • Rich formatting with emojis and clear sections

โšก Performance Examples:

  • Single recipe: ~150-300ms (cached responses faster)

  • Batch (3 cocktails): ~250-400ms (vs 900ms+ sequential)

  • Mixed batch (5 cocktails): ~300-500ms with parallel processing

  • Cache hit: <50ms instant response

๐ŸŽ›๏ธ Batch Control Parameters:

  • limit: 1-20 recipes (default: 10) - controls batch size

  • include_variations: Boolean - adds similar cocktails to results

Inputs: cocktail_id, cocktail_name, cocktail_ids, cocktail_names, +2 more. Returns: recipes, performance

smart_search_cocktailsA

๐Ÿš€ PREFERRED TOOL: Advanced cocktail search with intelligent batch processing and complete recipes.

๐ŸŽฏ BATCH PROCESSING SYSTEM:

  • High Performance: Parallel processing with 5-10x speed improvement

  • Smart Caching: Automatic caching for 70%+ faster repeated searches

  • Error Resilience: Individual failures don't break entire batch operations

  • Flexible Limits: Configure result count (default: 20, max: 50)

๐Ÿ“‹ Use Cases:

  • General searches: "gin cocktails", "winter drinks", "classic cocktails"

  • Similarity queries: "cocktails like Manhattan", "similar to Negroni"

  • Ingredient-based: "cocktails with bourbon", "drinks using Campari"

  • Flavor profiles: "bitter cocktails", "sweet drinks", "herbal spirits"

  • Complex filtering: combine ingredients, ABV ranges, glass types, methods

  • Batch comparisons: Multiple ingredient searches simultaneously

๐Ÿ”„ Batch Processing Examples:

  • Single search: {query: "Manhattan"} โ†’ Complete recipe + similar cocktails

  • Multi-ingredient: {ingredient: "gin", must_include: ["vermouth", "bitters"]}

  • Similarity batch: {similar_to: "Negroni", limit: 10} โ†’ 10 similar cocktails

  • Complex filter: {preferred_flavors: ["bitter"], abv_min: 25, limit: 15}

๐Ÿ“Š Response Format: Returns structured data with complete recipes including:

  • Ingredients with precise measurements in oz (auto-converted from ml)

  • Step-by-step preparation instructions

  • Cocktail specifications (ABV, glass, method, garnish)

  • Direct links to cocktail database pages

  • Performance metrics (processing time, cache hits)

  • Similar cocktail recommendations with full recipes

โšก Performance Features:

  • Parallel API processing for multiple results

  • Intelligent caching system with TTL management

  • Batch fetching of complete recipe details

  • Error isolation and fallback handling

Inputs: query, similar_to, similar_to_id, ingredient, +9 more. Returns: ๐ŸŽฏ Batch processing response with complete cocktail data and performance metrics

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelogโ€ฆ

    TDQS

    A4/5.0

    Scored across 3 tools

    Disambiguation4/5

    The three tools have distinct primary purposes: get_ingredient_info focuses on ingredient details and substitutions, get_recipe retrieves specific cocktail recipes, and smart_search_cocktails searches for cocktails based on various criteria. However, there is some functional overlap between get_recipe and smart_search_cocktails, as both can return complete recipes, which could cause minor confusion for an agent deciding which to use for recipe retrieval.

    Naming Consistency4/5

    The tool names follow a consistent verb_noun pattern (get_ingredient_info, get_recipe, smart_search_cocktails), which is clear and predictable. The only minor deviation is that smart_search_cocktails uses an adjective (smart) before the verb_noun structure, but this is still readable and maintains overall consistency.

    Tool Count5/5

    With three tools, this server is well-scoped for a bar assistant domain, covering key areas like ingredient information, recipe retrieval, and cocktail search. Each tool serves a distinct and valuable function, and the count is neither too sparse nor overwhelming, making it easy for an agent to navigate and use effectively.

    Completeness3/5

    The toolset covers core functionalities for a bar assistant, such as ingredient lookup, recipe fetching, and cocktail search. However, there are notable gaps in the lifecycle coverage; for example, there are no tools for creating, updating, or managing cocktails or ingredients (e.g., add_recipe, update_ingredient), which limits the server's utility for more interactive or creative tasks beyond retrieval and search.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      C
      maintenance
      Provides cocktail recommendations using a Retrieval-Augmented Generation (RAG) pipeline powered by LangChain, FAISS, and Groq. It enables users to search for cocktail recipes and receive personalized drink suggestions through natural language.
      6
      -
    • F
      license
      A
      quality
      D
      maintenance
      Enables natural language search for cocktail recipes and ingredient information through TheCocktailDB API. Supports searching by cocktail name, ingredient, category, or alcohol content to discover recipes and recommendations.
      5
      -