---
name: food-jobs
description: "Step-by-step job recipes for the Food Facts MCP server: Look up a barcode; Search products; Compare products. Use when the user asks things like \"What is in the product with barcode 3017620422003?\" or \"Find oat milks sold in France.\""
license: MIT
---

# Food Facts jobs

Each section below walks through one job with the tools of the `food` MCP server. The same recipes are MCP prompts (prompts/list) on `https://api.openkrill.app/gateway/mcp`.
Connect the server at `https://food.openkrill.app/mcp` in your MCP client settings (remote streamable HTTP, no key), then call its tools as normal tool calls. They are not command line programs: Bash, curl and scripts cannot reach them.

## Look up a barcode

Job: Look up a barcode, with the Food Facts MCP server.

Use the `lookup_barcode` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "What is in the product with barcode 3017620422003?"

Steps:
1. Call `lookup_barcode` as a normal MCP tool call, with the request's values in the fields of its inputSchema. It needs `barcode`.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"barcode":"3017620422003"}`
4. For the request "What is in the product with barcode 3017620422003?", a correct answer looks like this: status ok with the name Nutella, brands, ingredients text, allergens (milk, nuts, soybeans), Nutri-Score, NOVA group 4, nutrition per 100 g, a data_quality_note and the Open Food Facts link, with the Open Food Facts attribution and the reminder to check the physical label.
5. Follow the server's own rules:
   - Use these tools for questions about packaged food products identified by barcode or name.
   - Call lookup_barcode for one product, search_products to find barcodes by name or brand, compare_products to rank up to five barcodes by sugars, salt or kcal per 100 g, and check_allergens to see whether a product's listed allergens include ones the user wants to avoid.
   - The data is crowd-sourced: always mention the data_quality_note and tell the user to check the physical label, and never describe an allergen check as a guarantee that a product is safe.
   - Credit Open Food Facts.
   - Do not give dietary or medical advice.
   - Open Food Facts has no recall data: say so when asked about a recall.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `lookup_barcode`).

## Search products

Job: Search products, with the Food Facts MCP server.

Use the `search_products` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Find oat milks sold in France."

Steps:
1. Call `search_products` as a normal MCP tool call, with the request's values in the fields of its inputSchema. It needs `query`.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"query":"oat milk","country":"France"}`
4. For the request "Find oat milks sold in France.", a correct answer looks like this: status ok and a list of products, each with barcode, name, brands and Nutri-Score, taken from entries sold in France.
5. Follow the server's own rules:
   - Use these tools for questions about packaged food products identified by barcode or name.
   - Call lookup_barcode for one product, search_products to find barcodes by name or brand, compare_products to rank up to five barcodes by sugars, salt or kcal per 100 g, and check_allergens to see whether a product's listed allergens include ones the user wants to avoid.
   - The data is crowd-sourced: always mention the data_quality_note and tell the user to check the physical label, and never describe an allergen check as a guarantee that a product is safe.
   - Credit Open Food Facts.
   - Do not give dietary or medical advice.
   - Open Food Facts has no recall data: say so when asked about a recall.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `search_products`).

## Compare products

Job: Compare products, with the Food Facts MCP server.

Use the `compare_products` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Which has less sugar per 100 g: 3017620422003 or 5449000000996?"

Steps:
1. Call `compare_products` as a normal MCP tool call, with the request's values in the fields of its inputSchema. It needs `barcodes`.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"barcodes":["3017620422003","5449000000996"],"field":"sugars"}`
4. For the request "Which has less sugar per 100 g: 3017620422003 or 5449000000996?", a correct answer looks like this: status ok and a ranking from lowest to highest sugars per 100 g with each product's name, value and Nutri-Score, stating the basis is per 100 g.
5. Follow the server's own rules:
   - Use these tools for questions about packaged food products identified by barcode or name.
   - Call lookup_barcode for one product, search_products to find barcodes by name or brand, compare_products to rank up to five barcodes by sugars, salt or kcal per 100 g, and check_allergens to see whether a product's listed allergens include ones the user wants to avoid.
   - The data is crowd-sourced: always mention the data_quality_note and tell the user to check the physical label, and never describe an allergen check as a guarantee that a product is safe.
   - Credit Open Food Facts.
   - Do not give dietary or medical advice.
   - Open Food Facts has no recall data: say so when asked about a recall.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `compare_products`).