Restaurant Food Inventory Template with Purchase Units

Use a restaurant food inventory template that gives every ingredient one base unit, adds full packs and loose stock in that unit, and rounds the shortage up to whole purchase packs. Track mass in grams, volume in milliliters, or count in eaches. Do not convert grams to milliliters unless you have measured data for that specific ingredient.

Download the Excel workbook · Count lines (CSV) · Count sessions (CSV) · Purchase units (CSV) · Example rows (CSV) · Application field specification (JSON)

The synthetic examples are independently checkable: flour is 1 × 25,000 g + 4,000 g = 29,000 g on hand against a 35,000 g target, so order one 25,000 g pack. Milk is 2 × 6,000 mL + 500 mL = 12,500 mL against a 14,000 mL target, so order one case.

Give every ingredient one base unit

Create a stable ingredient_id and choose its dimension and base unit when the ingredient is added: mass with g, volume with mL, or count with each. Set the purchase pack size in that same base unit. A flour bag might contain 25,000 g; a milk case might contain 6,000 mL; a carton might contain 12 each.

Keep each ingredient’s units consistent through its count, target, and order calculation. A mass measurement and a volume measurement cannot be added directly. Do not treat 1 gram as 1 milliliter. If a recipe or supplier provides ingredient-specific density or measured conversion data, store that evidence and conversion rule explicitly; otherwise keep the original dimensions separate.

Convert packs and loose stock to base units

AI-generated editorial illustration.

Record the number of full purchase packs and any loose base units left outside full packs. Calculate:

On hand = full packs × base units per pack + loose base units

For example, one synthetic 25,000 g flour bag and 4,000 g loose flour give 1 × 25,000 + 4,000 = 29,000 g. Two 6,000 mL milk cases and 500 mL loose milk give 12,500 mL. Two 12-count egg cartons and three loose eggs give 27 each.

Use positive pack sizes and nonnegative whole full-pack counts. For counted items such as eggs, loose quantity and targets must also be whole numbers. Mass and volume can use decimals. Keep loose quantity in the base unit, not “partial case” or an ambiguous fraction, so count staff can enter the amount they actually measured.

Round the order up to whole packs

Set a target on-hand quantity per ingredient and calculate the shortage after the count. The order quantity is:

Order packs = ceil(max(0, target − on hand) ÷ base units per pack)

For each row, the deficit is greater than zero and no larger than one pack, so the ceiling is one pack. If on-hand stock already meets or exceeds the target, max(0, target − on hand) is zero and the order is zero packs. The formula does not include incoming orders. Check open purchase requests before sending another, and count arrivals as stock only when received. The calculated packs are a recommendation, not a placed order.

Use the synthetic sample files

sample.csv contains one row per ingredient with its dimension, pack size, count, target, and calculated order. All ingredients and quantities are invented. Check each result within its own unit: flour stays in g, milk stays in mL, and eggs stay in each. The sample’s ordered stock after replenishment would be 54,000 g of flour, 18,500 mL of milk, and 39 eggs if the full recommended pack arrives and no stock changes first.

Use purchase_units.csv to keep pack definitions separate from count rows. In a spreadsheet, validate that pack size is positive, full packs are whole numbers, and loose units use the selected base unit. Recount an ingredient if its dimension, supplier pack, or loose quantity is uncertain rather than silently converting between mass and volume.

Move food counts into an app

These CSVs work as editable templates while a restaurant refines its counting process. Preserve ingredient, purchase-unit, location, session, and line IDs if you move the records to a workflow app. Atoms offers Atoms Cloud and a connected Supabase project as alternative backend choices; connecting Supabase alone does not create inventory forms, validation, or durable saves. Once an app is built, verify that a saved count remains after reload and that another role sees only its permitted records. Atoms Supabase connection guide

For a one-time spreadsheet migration, see Spreadsheet to web app. To plan a dedicated workflow, visit the Atoms AI app builder.

Build a restaurant food inventory app in Atoms

Copy this complete prompt into your Atoms project chat:

Build a restaurant food inventory and purchasing app for a restaurant manager. Use one persistent backend: Atoms Cloud or a connected Supabase project. Keep editable CSV/spreadsheet templates available before the app is built.

Create stable-ID tables: ingredients (ingredient_id, name, dimension, base_unit, target_on_hand_base_units); purchase_units (purchase_unit_id, ingredient_id, pack_name, base_units_per_pack, supplier_reference); locations (location_id, name); count_sessions (session_id, location_id, counted_at, counted_by); count_lines (line_id, session_id, ingredient_id, purchase_unit_id, full_packs_on_hand, loose_base_units, on_hand_base_units); and purchase_requests (request_id, ingredient_id, purchase_unit_id, target_base_units, order_packs, status). Each ingredient has exactly one dimension and base unit: mass/g, volume/mL, or count/each. Never convert mass to volume without ingredient-specific measured data. Store pack size in that ingredient’s base unit.

Validate positive pack sizes and whole nonnegative full-pack counts. For count/each ingredients require whole loose quantities, pack sizes and targets; mass/volume can use nonnegative decimal loose quantities and targets. Calculate on-hand as full packs × base units per pack + loose base units. Calculate order packs as ceil(max(0, target − on-hand) / base units per pack). This formula excludes incoming orders; show open purchase requests beside the recommendation for manager review and require recorded receipt before increasing stock. Keep ingredients in separate dimensions; do not compare or add g, mL, and each. Preserve source IDs, count history, purchase requests, and deduplicated imports.

Use clearly labeled synthetic rows: flour uses g with a 25,000 g pack, 1 full pack and 4,000 g loose, target 35,000 g, order 1 pack; milk uses mL with a 6,000 mL case, 2 full cases and 500 mL loose, target 14,000 mL, order 1 case; eggs use each with a 12-count carton, 2 full cartons and 3 loose, target 32 each, order 1 carton. Expected on-hand values are 29,000 g, 12,500 mL, and 27 each; each order rounds up to a whole pack and never goes below zero.

Define server-side roles for organization admin, location manager, and count staff. Enforce organization and assigned-location row access, permitted fields and exports on the server. Do not rely on hidden controls. Validate units, pack references, IDs, quantities, target, and order calculation on the server. Do not assume POS, supplier, or purchasing API integration. Save one harmless edit, reload, and verify persistence; test a separate role and report each uncompleted check.