Model Context Protocol

Nutrition for your AI agent,
not from it.

A free MCP server that gives your AI agent USDA-grounded nutrition data. Stop letting it guess at your calories.

Connect Read the docs

Why this exists

LLMs generate text; they don't retrieve facts. Even with USDA data in their training mix, the numbers they produce are interpolated guesses, not lookups.

01
The problem

Your AI is making up macros.

An MDPI Nutrients study (2025) found GPT-4 underestimated calories by 36%, fat by 48%, and sodium by 53% in meal-photo nutrient estimation.

NutriBench reports 80% of LLM nutrition errors are carb predictions; the best model scored 51% accuracy.

Repeated queries returned different answers in a chatbot consistency study.

02
The fix

USDA data. Deterministic math.

An LLM parses your ingredients at temperature 0. The nutrients are then looked up in USDA FoodData Central (2M foods) and totaled by deterministic code. Same input, same output, every time. The model never invents a number; it reads one.

03
The signal

Confidence scores per ingredient.

Every ingredient comes back with match and conversion confidence between 0 and 1. Low scores flag fuzzy USDA matches or ambiguous portions. You know when to trust the number and when to double-check.

What you can do

Plug NutrientAPI into your AI agent once, then use plain language.

Example
Plan a recipe
You Build me a keto-friendly dinner under 600 calories. Verify the macros.
Agent [example dish name]. Verified against USDA:
XXX
kcal
XX g
protein
XX g
carbs
XX g
fat
Confidence 0.XX
analyze_recipe via NutrientAPI
Example
Analyze a recipe
You Here is a recipe I found online. Give me an FDA-style nutrition label per serving.
Agent Per serving (X servings total): XXX kcal, XX g fat (X g saturated), XX g carbs (X g fiber, X g sugar), XX g protein, XXX mg sodium. Iron XX% DV, calcium X% DV.
Confidence 0.XX · "[ingredient]" matched at 0.XX
analyze_recipe via NutrientAPI
Example
Look up a food
You What's the nutrition in 1 cup of cooked quinoa?
Agent Per 1 cup cooked: XXX kcal, XX g protein, XX g carbs, XX g fat, X g fiber.
Confidence 0.XX
get_nutrition via NutrientAPI

Pricing

MCP traffic counts against the same plan quotas as direct API calls.

Free
$0
25 recipes/month
Best for trying it out
Pay-as-you-go
$0.05/recipe
25 free/month, then $0.05 each
Best for low or unpredictable volume
Pro
$149/mo
10,000 recipes included
Best for daily use across many sessions
Max
$249/mo
50,000 recipes included
Best for production agents at scale

Works with Claude, ChatGPT, and any MCP-compatible client.

USDA FoodData Central · 2M foods OAuth 2.1 with PKCE Open MCP standard