HowToCook-mcp: Derivatives and Integrations of the HowToCook Repository

HowToCook-mcp is an ecosystem of AI-driven extensions that transform the Anduin2017/HowToCook recipe database into interactive personal chef assistants, available in both JavaScript and Python implementations.

The open-source cookbook Anduin2017/HowToCook maintains a structured collection of markdown-based recipes. While the core repository focuses on static documentation, its derivatives and integrations—specifically the HowToCook-mcp ecosystem—extend this content into dynamic, conversational interfaces. These tools parse the repository's markdown files to provide intelligent meal planning capabilities.

What Is HowToCook-mcp?

HowToCook-mcp represents the official community-maintained forks that implement the "Model Context Protocol" (MCP) pattern to expose HowToCook content through AI assistants. The project recognizes two primary implementations that consume the core repository's recipe data.

HowToCook-mcp (JavaScript/Node)

The original HowToCook-mcp is a Node.js-based implementation that exposes the recipe database through a JavaScript API. According to the repository's README.md at line 455, this derivative "turns an AI assistant into a personal chef, helping you plan breakfast, lunch and dinner for the whole day."

This implementation parses markdown content from the dishes/, tips/, and starsystem/ directories to extract structured data including ingredient lists, cooking steps, and nutritional notes.

HowToCook-py-mcp (Python)

HowToCook-py-mcp provides a Python-native implementation of the same AI-assistant concept. Listed at line 456 in README.md, this derivative offers identical day-meal planning capabilities through a Python-friendly API.

The Python implementation wraps the markdown parsing logic in a distributable package, making it easy to embed HowToCook content into custom scripts, data pipelines, or larger Python applications.

How HowToCook-mcp Integrates with the Core Repository

The HowToCook-mcp ecosystem does not duplicate recipe content. Instead, it implements a read-only integration pattern that consumes the original repository's structured markdown.

Content Source Directories

The derivatives specifically target three content directories within Anduin2017/HowToCook:

• dishes/ – Contains individual recipe markdown files organized by category
• tips/ – Stores cooking techniques and general kitchen advice
• starsystem/ – Holds advanced or rated recipes with structured metadata

Documentation and Template Propagation

The existence of these derivatives is documented in the project's central README.md file, ensuring visibility to all repository visitors. Furthermore, the integration is reinforced through the .github/templates/readme_template.md file, which automatically includes references to HowToCook-mcp and HowToCook-py-mcp in any new forks of the repository.

This template-based approach ensures that the ecosystem of derivatives remains discoverable regardless of how many community forks exist.

Using HowToCook-mcp: Code Examples

Both implementations expose a similar API pattern centered on the planDay (JavaScript) or plan_day (Python) method, which generates full-day meal recommendations based on dietary constraints and caloric targets.

JavaScript/Node Implementation

// Import the AI-assistant client provided by the HowToCook-mcp repo
const CookAssistant = require('howtocook-mcp');

// Initialise with an optional OpenAI key (handled internally)
const assistant = new CookAssistant();

// Ask for a full-day menu
assistant.planDay({
  dietary: 'vegetarian',
  calories: 1800,
}).then(menu => console.log('Your personalized menu:\n', menu));

Python Implementation

from howtocook_py_mcp import CookAssistant

assistant = CookAssistant()

# Generate a three-meal plan for a low-carb diet

menu = assistant.plan_day(diet='low-carb', target_kcal=1500)

print("Today's menu:")
for meal, recipe in menu.items():
    print(f"{meal.title()}: {recipe['title']} – {recipe['link']}")

Both examples demonstrate the core integration pattern: initialization loads the parsed markdown data from the source directories, the planning method applies AI-driven selection logic, and the returned objects contain direct references to the original recipe files in the dishes/ directory.

Summary

• HowToCook-mcp is the official ecosystem of AI-driven derivatives extending the Anduin2017/HowToCook repository.
• Two primary implementations exist: HowToCook-mcp (JavaScript/Node) and HowToCook-py-mcp (Python).
• Both integrations consume markdown content from the dishes/, tips/, and starsystem/ directories without duplicating data.
• The derivatives are documented in the main README.md (lines 455-456) and propagated through .github/templates/readme_template.md.
• Core functionality centers on the planDay/plan_day methods for generating personalized meal plans based on dietary constraints.

Frequently Asked Questions

What is HowToCook-mcp and how does it relate to the main repository?

HowToCook-mcp is a community-maintained ecosystem of AI assistants that transform the static recipe collection in Anduin2017/HowToCook into interactive meal planning tools. Rather than forking the content, these derivatives integrate directly with the original repository's markdown files in the dishes/ and tips/ directories, treating the main repo as a read-only data source.

What are the differences between HowToCook-mcp and HowToCook-py-mcp?

HowToCook-mcp is the original JavaScript/Node.js implementation, while HowToCook-py-mcp provides a Python-native alternative. Both offer identical core functionality—specifically the planDay (JS) and plan_day (Python) methods for generating daily meal plans—but target different runtime environments. The Python version is packaged for easy integration into data science workflows, whereas the Node version fits JavaScript application stacks.

How do these derivatives access the recipe data without duplicating it?

The integrations use a read-only consumption pattern. They parse the markdown files located in dishes/, tips/, and starsystem/ directly from the source repository. This approach ensures that any updates to recipes in the main Anduin2017/HowToCook repo are immediately reflected in the AI assistants without requiring manual synchronization or content duplication.

Where are these integrations documented in the source code?

The derivatives are officially listed in the main README.md file at lines 455 and 456, which describe HowToCook-mcp and HowToCook-py-mcp respectively. Additionally, these references are hardcoded into the .github/templates/readme_template.md file, ensuring that any new forks of the repository automatically inherit documentation pointing to these ecosystem projects.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}