goodbook-mcp
# Goodbook MCP Server
A modern Model Context Protocol (MCP) server that works exclusively with PDF files containing food service standards and preparation guidelines. Built following MCP best practices with robust input validation and error handling.
## Overview
The Goodbook MCP server allows AI assistants to query and retrieve information from food service documentation, including:
- Food preparation standards
- Cooking guidelines and methods
- Food safety information
- Recipe standards
- Hygiene and sanitation requirements
## Modern Architecture
This server follows current MCP best practices:
- **Zod input validation** for all tool parameters
- **Proper STDIO usage** with `console.error` for logging
- **Robust error handling** with detailed error messages
- **Optimized tool definitions** with automatic schema conversion
- **Comprehensive synonym support** for Russian cooking terms
- **Clean separation of concerns** between parsing, validation, and execution
## Features
### Available Tools
1. **search_food_standards**: Search for specific food preparation standards, recipes, or cooking guidelines
2. **get_cooking_guidelines**: Get cooking guidelines and standards for specific dishes or cooking methods
3. **list_sections**: List all available sections in the food service standards document
4. **get_section_content**: Get content from a specific section of the food standards document
5. **get_food_safety_info**: Get food safety information and hygiene standards
6. **find_recipe_standards**: Find standardized recipes and preparation methods for specific dishes
## Installation
1. Ensure you have Node.js 18+ installed
2. Clone or download this project
3. Install dependencies:
```bash
npm install
```
4. **REQUIRED:** Place your PDF file named `menu.pdf` in the root directory
## Usage
### Running the Server
```bash
npm start
```
The server will start and listen for MCP protocol messages via stdio.
### Development Mode
```bash
npm run dev
```
This will start the server with auto-reload on file changes.
### MCP Client Configuration
To use this server with an MCP-compatible client, add the following configuration:
#### Claude Desktop Configuration
Add to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"goodbook": {
"command": "node",
"args": ["c:/указываем-полный-путь-к-файлу/mcp/src/index.js"],
"env": {}
}
}
}
```
#### Other MCP Clients
Configure your MCP client to run:
```bash
node src/index.js
```
## Tool Usage Examples
### Search for Food Standards
```
Use the search_food_standards tool to find information about temperature requirements:
- query: "температура приготовления"
- section: (optional) specify a particular section
```
### Get Cooking Guidelines
```
Use get_cooking_guidelines to get preparation methods:
- dish_type: "паста"
- section: (optional)
```
### Food Safety Information
```
Use get_food_safety_info for safety guidelines:
- topic: "хранение мяса"
```
## File Structure
```
goodbook-mcp-server/
├── src/
│ ├── index.js # Main MCP server
│ ├── tools.js # Tool implementations
│ └── pdf-parser.js # PDF parsing logic
├── menu.pdf # Source PDF document
├── package.json # Project configuration
└── README.md # This file
```
## Requirements
- Node.js 18.0.0 or higher
- PDF document (`menu.pdf`) containing food service standards
- MCP-compatible client (like Claude Desktop)
## Error Handling
The server includes comprehensive error handling for:
- PDF parsing failures
- Missing sections
- Invalid search queries
- Tool execution errors
All errors are returned as structured responses to the MCP client.
## Development
### Adding New Tools
1. Add tool definition to `getToolDefinitions()` in `tools.js`
2. Implement handler method in `GoodbookTools` class
3. Add case to `handleTool()` switch statement
### Modifying PDF Parsing
Edit the `PDFParser` class in `pdf-parser.js` to:
- Change section detection patterns
- Modify content extraction logic
- Add new search capabilities
## License
MIT License
## Support
For issues or questions, please check the error logs when running the server. The server outputs detailed logging information to help diagnose problems.TDQS
Scored across 6 tools
There is significant overlap between tools like find_recipe_standards, get_cooking_guidelines, and search_food_standards, which all seem to retrieve similar types of information about recipes, guidelines, and standards. However, the descriptions provide some differentiation, such as get_food_safety_info focusing on safety/hygiene and list_sections/get_section_content being more structural. The overlapping purposes could lead to confusion when selecting the right tool for a query.
The tool names follow a consistent verb_noun pattern with snake_case throughout, such as find_recipe_standards and get_cooking_guidelines. All tools start with verbs like 'find', 'get', 'list', or 'search', which is predictable and readable. There are no deviations in naming conventions, making it easy to parse the tool set at a glance.
With 6 tools, the count is well-scoped for a server focused on food service standards and recipes. This number is neither too thin nor too heavy, allowing for a focused set of operations without overwhelming an agent. Each tool appears to serve a distinct purpose within the domain, making the count appropriate for the server's apparent scope.
The tool set covers core operations for accessing food standards, recipes, safety info, and document structure, providing good coverage for retrieval and search within the domain. However, there are minor gaps, such as no tools for updating or managing standards (e.g., create or modify operations), which might limit agents in more dynamic workflows. Overall, it supports essential querying tasks effectively.