Restaurant data workflows
How to Get Toast Data into Google Sheets
You want your Toast numbers in a spreadsheet, not another dashboard. Start by getting one weekly report into Google Sheets with the right dates, the right sales figure and a row you can update next week.
Feast's free restaurant data connector lets your AI read the available imported POS records. Ask for a tab-separated report, paste it into Sheets and compare it with Toast. Once that report is right, you can build a recurring refresh around the same layout.
- Toast
Your restaurant's records
- Feast
Connection, imports and storage
- Your AI
Prepare the report to copy
- Google Sheets
The spreadsheet you work in
Only need the data once? Start with a CSV
Toast supports downloading selected reports as CSV files. Export the report and date range you need, then import the file into Google Sheets. Available columns and export options depend on the Toast report.
That is often enough for a one-off analysis. The workflow below uses Feast and your AI to prepare a consistent weekly summary. You paste the first report yourself; automatic updates are a separate next step.
Toast's report-export instructionsStart with one row per location and week
A weekly summary does not need every check or your guests' contact details. Start with the location, week, order amount, number of orders and average order amount. Keep the dates, timezone and meaning of the amount alongside the report so you can compare it with Toast.
| Location | Week | Order amount | Orders | Average order | Data status |
|---|---|---|---|---|---|
| Example location | Oct 5–11, 2026 | $7,350 | 210 | $35 | Provisional |
This example uses America/New_York and a consistent order-amount definition. Provisional means the imported records have not been verified as complete. The sheet should also retain the last successful report time and the source freshness when known.
1. Connect Toast and select the right location
Start the free Feast setup, choose Toast and connect the restaurant you want to report on. Complete the required permissions and confirm that the history you need is available. Your AI needs an authorized Feast connection too; the setup guide explains that part.
Ask the AI to identify your restaurant and location before reading its data. Feast's listOrganizations tool finds the organizations your account can access. describeData then shows the available fields and query instructions. This avoids building a report around a field your connection does not contain.
2. Choose the sales figure before building the report
Choose the Toast report you want the sheet to agree with. Confirm whether its amount includes tax, tips, discounts or refunds, and which order statuses it includes. Have the AI inspect the available fields rather than assume every POS uses the same definitions.
For a weekly report, set a Monday-through-Sunday period in the restaurant's timezone. The week, location and amount definition must match on both sides. Order amount is not automatically net sales, marketing-attributed revenue or profit.
Example request
Find the authorized Toast restaurant and location. Show the available order amounts, status fields, dates and timezone. Help me match the amount and included orders to the Toast report I use. Do not guess a net-sales definition or assume another POS's status values.3. Prepare one complete weekly result
Use queryData to read the documented fields. Ask for weekly totals where the source supports them. If the response has a nextCursor, continue with that cursor until every result page is included. The first page of orders is not necessarily the whole week.
Keep unavailable amounts and incomplete imports visible. A successful query means a query ran, not that all Toast records have arrived. Compare the result with the matching Toast report before treating the week as complete.
$7,350 ÷ 210 orders = $35 per order
4. Paste the report into a blank Google Sheets tab
Ask your AI for tab-separated text with the header and result on separate lines. Open a blank tab in Google Sheets, select A1 once and paste the block. Each heading should have its own column, with the report on the next row. Use your own result for your restaurant; the practice block below contains only fictional data.
If everything lands in one column, select the pasted cells and use Data, then Split text to columns. Let Sheets detect the separator. Check that the amount and order columns are numbers rather than dollar-formatted text before adding formulas.
Keep the restaurant, source and timezone in a note above or beside your report. Use the location and week start to find its row. When you refresh the same week, replace that row's values; add a new row only for a different location or week.
Example request
Return the weekly report as tab-separated text I can paste into Google Sheets. Include the header once, then one row per location and week. Use these columns: Location, Week start, Week end, Order amount (USD), Orders, Average order (USD), Data status. Keep amounts numeric, without currency symbols or thousands separators. Keep dates as YYYY-MM-DD. Put the restaurant, source, timezone and amount definition outside the block. Do not include guest details or write to a spreadsheet.5. Run it again. The same week should still have one row
Run the report manually, compare it with Toast, then run it a second time. If nothing changed, the values should stay the same. If more records have arrived, update that week's existing row rather than adding a second version to the totals.
6. Want automatic refreshes? Plan that as the next step
The copy-and-paste workflow above is manual. To automate it, choose a runner that can read Feast and write to Google Sheets, authorize the destination and arrange its schedule and execution host. A connected AI chat does not automatically become a scheduled job.
Give the builder the same restaurant, source, location and week-start combination as the row key. The automation needs to find an existing key and update its range, or append a row only when the key is absent. A Sheets append operation alone does not prevent duplicates.
Confirm which account will run the job, its timezone, how it keeps authorized access and where it reports a failure. A working scheduled runner is additional implementation, not something this page activates.
Choose how many recent weeks each run will revisit, because older totals can change as records arrive. Keep a last-successful-run time in the sheet. If access fails or imports are incomplete, mark the affected period for review rather than replacing it with zero sales. Any low-sales alert must wait until the data it needs has been checked.
Give your AI the complete task
Prepare a weekly Toast report from my authorized Feast data that I can paste into Google Sheets. First identify the restaurant, Toast source, location and available fields. Agree the amount definition, timezone, week boundaries and how to check import coverage. Read the full weekly result, following every returned nextCursor when needed. Return tab-separated text with this header: Location, Week start, Week end, Order amount (USD), Orders, Average order (USD), Data status. Use numeric amounts without currency symbols or thousands separators, and YYYY-MM-DD dates. Put the restaurant, source, timezone, amount definition and any missing-data notes outside the block. Exclude guest details. Do not write to a spreadsheet or enable an automation. Help me compare the result with Toast, then rerun the same week so I can update its existing row. Never treat missing data as zero sales.Build the report, not the POS infrastructure
Feast handles the supported POS connection, imports, hosting, storage and query tools, so you and your AI can build on data that is already available. Your report does not need to request every answer directly from Toast; provider limits still apply to imports.
The Sheets connection, recurring job and any custom alerts need their own setup. Bring the workflow you want to your free setup call so the Feast team can help you work through those choices.