Google Sheets Sync — Reliable Import & Export
Pricing
Pay per usage
Reliable Google Sheets import & export: send Apify datasets or JSON to a sheet (append, replace, upsert with merge, dedupe) or import a sheet to a dataset. Auto-grows the grid, handles schema drift, retries rate limits, clear error messages.
"Export" writes rows into a Google Sheet. "Import" reads a sheet range and pushes each row as a dataset item.
Paste the full JSON key file downloaded for a Google Cloud service account with the Sheets API enabled. Share the target spreadsheet with the service account's client_email as Editor first — see the README setup guide. Leave blank when dryRun is enabled.
The ID from the sheet's URL: docs.google.com/spreadsheets/d/<THIS_PART>/edit. Not required when dryRun is enabled.
Name of the worksheet tab to read from or write to.
How to write rows in export mode. Append: add rows after existing data. Replace: clear the sheet and write fresh. Upsert: update rows matching Key column(s), append the rest. Dedupe: like append, but skip rows whose key already exists in the sheet or earlier in the same run.
Column/field name(s) that uniquely identify a row. Required for Upsert and Dedupe export modes.
[]Only used when exportMode is Upsert. Merge (default): a matched row keeps its existing cell values for any column the new row doesn't include, so a partial update never blanks out other columns. Replace: a matched row is overwritten entirely, exactly like the old behavior -- any column the new row omits is cleared.
Export rows from this dataset ID instead of the current run's default dataset. Leave blank to use sourceRunId, inputRows, or the default dataset (checked in that order).
Export rows from the default dataset of this earlier Actor run ID. Used only if sourceDatasetId is blank.
Export exactly these JSON objects instead of reading a dataset. Handy for testing or for small ad-hoc pushes. Used only if sourceDatasetId and sourceRunId are both blank.
[]Turn nested objects into dot-separated columns, e.g. address.city. Turn off to JSON-stringify nested values into a single cell instead.
Arrays of plain values (strings/numbers/booleans) are joined into one cell with this delimiter. Arrays of objects are always JSON-stringified.
Range to read in import mode, e.g. A1:F1000. The sheetName above is used as the tab unless you include a tab name here yourself.
Row within the import range that holds column headers. Rows above it are ignored; rows below become dataset items keyed by these headers.
How Google Sheets interprets values you write. RAW (default): stored exactly as given -- a number like 54.07 is always stored as that number, regardless of the sheet's locale (e.g. a comma-decimal locale). USER_ENTERED: parsed as if a person typed it into the UI -- needed for a cell that should become a formula (e.g. "=A1+B1"), but risks locale-dependent reinterpretation of plain numeric/date-like strings.
If the named sheet tab does not exist yet in the spreadsheet, create it instead of failing (export mode only).
Store write progress in the key-value store so that if this run is retried (same spreadsheet, sheet, and mode), it continues from where it left off instead of rewriting or re-appending already-written rows.
Self-imposed ceiling on cells written per Sheets API request, well under Google's per-request limits, to keep requests small and retry-friendly.
Self-imposed ceiling on the JSON payload size per Sheets API request, kept under Google's 10 MB request limit.
How many times to retry a single Sheets API call after a 429 (rate limit) or 5xx (server error) response, with exponential backoff and jitter, before giving up.
Run the full pipeline (read input, flatten/coerce rows, batch, resume-check) against an in-memory fake sheet instead of calling Google. Useful for testing without credentials. The dataset gets a summary item describing what would have been written/read.