Cookbook MCP Server

The Cookbook MCP Server is a lightweight, domain‑specific tool that lets AI assistants fetch and present cooking recipes from an offline or local repository. By exposing two core tools— and —the server turns a static collection of recipe files into an interactive API that can be queried on demand. This solves the problem of having to manually parse and search through a large directory of Markdown or text files, allowing developers to focus on higher‑level conversational flows rather than file handling logic.

At its heart, the server reads a cookbook directory (typically cloned from a public repository such as HowToCook) and exposes two endpoints: one that lists every dish available, and another that returns the full recipe for a requested item. For developers building AI assistants that need culinary knowledge, this means instant access to accurate, up‑to‑date cooking instructions without relying on external paid APIs. The server can be launched either as a standard HTTP service or via Server‑Sent Events (SSE), giving flexibility in how clients consume updates and responses.

Key capabilities include:

• Simple discovery – returns a structured list of dish names, enabling quick menu generation or autocomplete suggestions.
• Targeted retrieval – fetches a specific dish’s instructions, ingredients, and preparation steps in plain text or JSON format.
• Local data privacy – Since the cookbook is stored locally, sensitive dietary preferences or proprietary recipes remain on‑premises.
• SSE support – When configured, the server streams responses in real time, which is useful for long recipes or step‑by‑step guidance.

Typical use cases span a range of scenarios: an AI kitchen assistant that walks users through cooking steps, a chatbot that recommends meals based on available ingredients, or an educational tool that teaches culinary techniques. Developers can integrate the server into existing MCP workflows by adding a single configuration entry pointing to the executable, then invoking the tools from prompts or tool calls within their assistant. The design is intentionally minimalistic yet powerful, making it easy to extend with additional tools (e.g., nutritional analysis) without changing the core architecture.

What sets this server apart is its zero‑friction integration with MCP clients and its focus on a niche yet widely useful domain. By turning static recipe files into first‑class conversational resources, it empowers developers to build richer, more engaging culinary experiences while keeping data control in their hands.