Foodblock Mcp
Wire natural-language food descriptions into agents and apps that need structured nutrition or catalog fields.
Overview
FoodBlock MCP is a MCP server for the Build phase that converts English food descriptions into structured food data for AI agents.
What is this MCP server?
- Turn plain-English food descriptions into structured food records via MCP tools
- Stdio transport npm package foodblock-mcp (registry version 0.5.1)
- Optional FOODBLOCK_URL env for hosted FoodBlock vs standalone local mode
- Designed for any AI agent workflow that needs consistent food entities, not one-off prompts
- Server schema version 0.5.1
- Single npm stdio package identifier foodblock-mcp
- Optional FOODBLOCK_URL environment variable for non-standalone mode
What problem does it solve?
Agents hallucinate nutrition facts and inconsistent food labels when you rely on free-form LLM output alone.
Who is it for?
Solo builders shipping food, fitness, or grocery agents who want structured food entities without building a custom parser first.
Skip if: Teams that do not model food at all or already own a locked-in nutrition API with no need for MCP exposure.
What do I get? / Deliverables
Your agent can fetch normalized food structures through MCP so features like logging, search, and recommendations use one data shape.
- MCP tools that return structured food data from English descriptions
- Agent workflows that reuse consistent food records across sessions
- Optional linkage to a hosted FoodBlock instance via FOODBLOCK_URL
Recommended MCP Servers
Journey fit
Food data normalization belongs in Build when you connect agents to real domain APIs instead of hand-parsing user meal text. Integrations is the right shelf because the server exposes MCP tools over stdio so Claude Code or Cursor can call FoodBlock during implementation.
How it compares
Domain food-data MCP integration, not a general web-scraping or recipe-generation skill.
Common Questions / FAQ
Who is FoodBlock MCP for?
Indie developers and agent builders who need dependable structured food data from natural language inside Claude Code, Cursor, or similar MCP clients.
When should I use FoodBlock MCP?
Use it during Build when you integrate meal logging, nutrition search, or catalog features and want tools the model can call instead of guessing food fields.
How do I add FoodBlock MCP to my agent?
Register the npm stdio package foodblock-mcp in your MCP config, set FOODBLOCK_URL only if you use a remote FoodBlock server, then restart the agent host.