Bar Assistant MCP Server
# Bar Assistant MCP Server
A high-performance Model Context Protocol (MCP) server that provides intelligent cocktail search and recipe retrieval from [Bar Assistant](https://github.com/karlomikus/bar-assistant) instances.
## Features
- 🔍 **Smart Cocktail Search** - Natural language queries with similarity matching
- 📖 **Complete Recipes** - Detailed ingredients, instructions, and specifications
- 🎯 **Similarity Engine** - Discover cocktails similar to your favorites
- 🧪 **Ingredient Analysis** - Find cocktails by ingredients, flavors, and characteristics
- ⚡ **Batch Processing** - Retrieve multiple recipes simultaneously (5-10x faster)
- 💾 **Smart Caching** - Intelligent caching with 70%+ hit rate
- 🔧 **Advanced Filtering** - ABV ranges, glassware, preparation methods, and more
## 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 → 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:**
```bash
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?"*TDQS
Scored across 3 tools
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.
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.
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.
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.