# MCP Servers with a search_recipes tool

Every MCP server Pod knows of that exposes a search_recipes tool.

8 servers, ordered by decision readiness and adoption — GitHub stars and npm downloads where available, then reported issue volume. Pod has connected to 7 of them; the rest are publisher-reported and unverified.

## [Dillr](/mcp/dillr)

Connect your Dillr kitchen to Claude so it knows what's in your pantry, who's eating, and what you've cooked before. Claude can plan meals around real constraints, save recipes to your library, build shopping lists, and update your inventory — all in one conversation. Everything Claude writes syncs instantly to the Dillr app for everyone in your household, so you can plan at your desk and cook from your phone.

23 tools reported — Live. Answered with 15 tools.

## [Heathy — Nutrition Data](/mcp/heathy-nutrition-data)

Heathy gives Claude nutrition data for 1,164 whole plant foods and 39 Ayurvedic herbs, drawn from 31 national food-composition tables — USDA, UK CoFID, India's IFCT 2017, ANSES-Ciqual, BLS, Fineli, Japan's MEXT and more.

Most nutrition sources give one number from one table. Heathy returns the number and its provenance: the median across the tables that have the food, how many sources contributed, which ones, and the full range between them. That lets Claude cite a figure rather than assert it, and say when the tables disagree. Guava vitamin C, for example, is 222 mg as the median of 13 tables, with values ranging from 68 to 273 mg.

Tools:
• lookup_food — full per-100 g profile (131 nutrients) with sources and spread
• search_foods — resolve names, including Hindi, Sanskrit and botanical aliases
• nutrient_ranking — "which foods are highest in iron", with a minimum-sources filter
• check_claim — which "source of" / "high in" nutrition claims the published data supports under GB/EU and US rules for a raw food, and where that depends on which national table was used (screening only, not a compliance sign-off)
• herb_monograph — Ayurvedic herbs with traditional use and research kept separate, and a safety section with cautions, pregnancy guidance and drug interactions
• search_recipes / get_recipe — original raw vegan recipes with computed nutrition
• health_topic — what to eat for a goal, with notes on what the evidence does and does not support

Read-only, no account needed. Informational only — not medical advice.

0 GitHub stars · 8 tools reported — Live. Answered with 8 tools.

## [Honeydew](/mcp/honeydew)

Connect your Honeydew cookbook to Claude. Search everything you've saved by ingredient, cuisine, or craving, browse your folders and tags, and pull up full ingredients, steps, notes, and nutrition for any recipe. Then put it to work: ask Claude to build a week of dinners from your favorites, or roll a few recipes into a single grocery list. Try "what can I make with chicken and spinach?", "plan three quick weeknight meals from my saved recipes," or "make a shopping list for these." Claude reads from your own collection — it never adds, edits, or deletes your recipes.

26 tools reported — Live, but requires authorization before it will list tools. Publisher lists 26 tools.

## [weReci](/mcp/wereci)

weReci is your own recipe collection — the dishes you've saved, the shelves
an agent curates for you, and the web of connections between ingredients,
techniques, and entries that grows as you cook. This connector lets Claude
read that collection directly, so you can ask about your own recipes the
same way you'd ask a friend who actually knows your kitchen.

Search your saved recipes by ingredient, cuisine, or half-remembered idea.
Browse the curated shelves weReci's grouping agent builds for you — themed
collections like "weeknight pasta" or "what's in season" — and ask Claude to
reshuffle them with a creative direction of your choosing. Explore the
connections graph that links your recipes to ingredients, techniques, and
the other entries in your collection, so you can trace how one dish relates
to another or find the thread between a recipe and the person who gave it to
you. If you share a joint cookbook with someone, Claude sees that shared
collection too, exactly as you do in the app.

weReci also keeps a personal memory layer — durable facts about your
preferences, dietary needs, and patterns, learned from how you actually use
the app. Claude can read that memory to personalize its answers instead of
asking you to repeat context every time.

The connection is read-only by default. The one optional write action —
emailing a recipe — has no recipient field anywhere in the tool: it can only
ever send to the verified email address on your own signed-in weReci
account, so there's no way for a misbehaving or prompt-injected agent to
mail anyone else. A second tool can regenerate your curated shelves on
request, but it only rearranges shelf themes — it never touches, edits, or
deletes a saved recipe.

Sign in with your existing weReci account through standard OAuth. Revoke
access anytime from weReci settings → Connected apps — nothing is readable
after you do.

10 tools reported — Live, but requires authorization before it will list tools. Publisher lists 10 tools.

## [Yokitup](/mcp/yokitup)

Yokitup is a restaurant inventory and operations management platform designed for restaurant operators and managers.

The Yokitup MCP server connects Claude to operational data so users can explore, analyze and better understand their restaurant performance using natural language.

Claude can access data such as products, recipes, food costs, inventories, stock losses and waste, stock transfers, and sales. It can combine information from different areas of the business to answer operational questions, identify trends and highlight anomalies.

For example, users can ask Claude to identify the products generating the highest losses, analyze food costs and margins, review inventory data, investigate stock movements, compare sales with operational data, or compare performance across locations.

The integration is particularly useful for restaurant groups and multi-site operators who need to analyze large amounts of operational data without manually exporting and combining reports.

Yokitup helps restaurant teams turn inventory, recipe, cost and sales data into actionable insights, with a particular focus on controlling food costs, understanding margins and improving operational performance.

31 tools reported — Live, but requires authorization before it will list tools. Publisher lists 31 tools.

## [GD](/mcp/gd)

Good Day (GD) is a curated, plant-forward recipe catalog for Claude. Ask for a recipe in plain language — "quick vegan dinner," "warm and spicy for a cold night" — and get ranked matches with times, diet type, and ratings, or browse by meal type, diet, and cooking time. Every recipe comes complete: ingredients with quantities and step-by-step instructions. GD's ~84 (and growing) recipes are vegan and vegetarian, strongest in quick savory meals. v1 is public and read-only, with one additive write: when nothing matches, Claude can log your request so future batches fill real gaps. Meal-planning and pantry features are planned for a future authenticated release.

4 tools reported — Did not respond when Pod last checked.

## [StewAI](/mcp/stewai)

StewAI helps Claude find and compare accessible AI workflows, estimate and run them, inspect declared results and settled billing, or build, test, qualify, and privately reuse a new recipe—without managing an API key.

14 tools reported — Did not respond when Pod last checked.

## [Stackcut by stackcut](/mcp/stackcut-by-stackcut)

Stackcut answers the money questions about software: is this subscription worth it, what's cheaper, which vendor fits what I'm building, and what does my whole stack cost.

Ask it about:
• One product you pay for. "Is there something cheaper than Typeform? We pay $100 a month." Stackcut compares a smaller plan, cheaper vendors and a build-it-yourself path, with first-year savings, setup time and upkeep.
• A whole stack. "Audit Notion, Asana, Loom and Zapier for a 12-person team." It asks for the few facts that decide the answer (seats, the features you use, usage), then shows what each change would save at your prices.
• Something you're building. "What will auth, transactional email and Postgres cost at 10,000 monthly users?" Or paste a product brief: it works out the jobs, leaves out what the brief rules out, prices each job at your scale and flags paid add-ons and budget overruns.
• A workflow change: billing shared across projects, moving hosting or email, cheaper AI model routing.

How it answers: prices come from Stackcut's catalog of vendor plans, with published limits, add-ons, a source link and the date each price was checked. A saving counts as proven only when every check passes. Until then Stackcut returns the open questions instead of a recommendation, and labels estimates as estimates.

What it doesn't do: it never cancels, buys or signs up for anything. No account, login or API key. Send names, numbers and prices only, never source code, secrets or invoices. This endpoint keeps no tool arguments or answers; it only counts calls.

16 tools reported — Publisher-reported; not verified by Pod.

## For agents

This page has a [Markdown](/mcp/tool/search-recipes.md) and a [JSON](/mcp/tool/search-recipes.json) twin. Pod is also queryable over MCP at `https://api.askpod.ai/mcp/read`.