# FoodXDevelopment/foodblock-mcp [Health: Active]

**Category:** 🛠️ Other Tools and Integrations  
**Repository:** https://github.com/FoodXDevelopment/foodblock  
**GitHub Stars:** 1  
**npm Downloads (last month):** 238  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/foodxdevelopment-foodblock-mcp

## Description
17 MCP tools for the food industry. Describe food in plain English and get structured, content-addressed data blocks back. Covers actors, places, ingredients, products, transforms, transfers, and observations. Works standalone with zero config via npx foodblock-mcp.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "foodblock-mcp": {
    "url": "https://api.example.com/foodblock"
  }
}
```

## Documentation

## What FoodXDevelopment/foodblock-mcp does

FoodXDevelopment/foodblock-mcp exposes FoodBlock capabilities through 17 MCP tools for representing food-industry information. It accepts natural-language descriptions such as a product listing, a farm description, a review, a purchase order, or a temperature reading, then returns structured blocks. The supported domain includes actors, places, substances such as ingredients and products, transformations, transfers, and observations.

The underlying model is intentionally small. Every block contains a dot-notated type, a JSON state object, and named references to other blocks. Examples include `actor.producer`, `substance.product`, `transform`, `transfer.order`, and `observe.review`. The state is schemaless, so applications can store fields appropriate to their operation without configuring a fixed database schema.

## How it works

FoodBlock derives each block's identity from a canonical representation of its type, state, and references, then hashes that representation with SHA-256. Identical content therefore produces the same identifier. References can connect a finished product to its ingredients, seller, source, or other related blocks, allowing provenance and supply-chain relationships to be represented as a graph.

Natural-language parsing uses local pattern matching and built-in vocabularies rather than an LLM or network API. The parser can identify common food-related intents, quantities such as prices, weights, volumes, temperatures, and ratings, as well as flags and relationships expressed in text. The broader FoodBlock API also supports creating, updating, validating, signing, resolving, and traversing blocks, but the available MCP tool surface should be checked before depending on a particular operation.

## Setup and configuration

FoodXDevelopment/foodblock-mcp runs standalone with zero configuration through the npm package command `npx foodblock-mcp`. No API key, remote account, database, or external service is required for the local natural-language functionality described in the README. The project also includes a separate sandbox that can be started with Node.js and exposes HTTP routes for testing block creation, lookup, filtering, batch synchronization, provenance chains, discovery, and natural-language input.

The repository separately documents the `foodblock` JavaScript package, which can be installed with npm for direct programmatic use. That package is distinct from the MCP server launch command and includes APIs for aliases, templates, URIs, signatures, tombstones, offline queues, and federation.

## Tools and capabilities

The material describes these supported capability areas:

- Parse food-related text into primary and related FoodBlocks.
- Represent actors, places, substances, transformations, transfers, and observations.
- Preserve relationships through hash-based references.
- Produce deterministic, content-addressed identifiers.
- Model provenance chains and supply-chain dependencies.
- Handle structured details such as price, weight, temperature, ratings, allergens, and food attributes when detected by the parser.
- Run locally without network calls or paid API credentials.

## Limitations and notes

The natural-language interface is based on regular-expression pattern matching and built-in vocabularies, not general-purpose language understanding. Its extraction behavior is therefore bounded by the food concepts and patterns implemented by the project. The state model accepts arbitrary valid JSON, but that flexibility does not imply automatic validation for every business-specific schema.

The README describes 17 MCP tools but does not provide their individual names in the supplied material. It also does not identify specific MCP clients, a software license, or a hosted service endpoint. Federation and remote resolution are documented as FoodBlock library features; they should not be assumed to be enabled by the standalone MCP process without checking the implementation.

## Using this FoodXDevelopment/foodblock-mcp MCP server
Always refer to the official documentation for the most accurate and up-to-date information on how to configure and run this server.

_Full upstream README: https://allmcps.com/mcp/foodxdevelopment-foodblock-mcp/readme_