Rows
ProductivityRead and write Rows spreadsheets from workflows. Agents read cell ranges, write values, and append rows, turning a Rows spreadsheet into a live input or output for any automation.
What This Integration Enables
Spreadsheets persist because they are the one interface every stakeholder already knows how to read. The failure mode was never the grid; it was the human courier work of getting data in and out of it. Rows leans into the grid, and this connector eliminates the courier: a Rows spreadsheet becomes a live input and output surface for FlowRunner agents. The agent does the fetching, computing, and syncing; the humans keep working in the grid they trust. - Read values from any A1-notation range, a block, a column, a summary row, to feed downstream steps - Write computed results back into specific ranges, replacing what was there deliberately - Append rows to the end of a table without disturbing existing data, the safe default for logs and syncs - Discover spreadsheet, page, and table IDs programmatically so flows target the right range as workbooks evolve - Keep the humans in the grid they know while agents handle every transfer in and out of it, on schedule or on event The data model rewards a moment of setup discipline. A spreadsheet contains pages, pages contain tables, and every read and write targets a table plus an A1 range, so flows that resolve their targets through Get Spreadsheet at run time survive workbook reorganizations that would silently break a hard-coded range. The agent re-finds its table; the flow keeps working through reorganizations that would break a hard-coded integration. The distinction between Append Rows and Write Cells is the connector's whole safety model, and it maps directly onto FlowRunner's [human-in-the-loop](/concepts/human-in-the-loop) discipline: additive writes flow, destructive writes ask.
Without FlowRunner
With FlowRunner
Use Case Scenarios
The cross-tool sync that keeps the ops sheet current
Records land in [Airtable](/integrations/airtable) all day. Each night, the agent pulls the day's new records, resolves the target table with Get Spreadsheet, and calls Append Rows to add them to the Rows tracking table. Nothing existing is touched, the sync is idempotent by run date, and operations reads one sheet instead of three tools. When the source adds a field, the agent flags the shape change instead of silently misaligning columns.
The KPI range that updates itself and shows its work
The leadership dashboard in Rows has a summary block driven by a fixed range. The agent computes the week's figures from source systems, calls Read Cells to capture the outgoing values, then Write Cells to place the new ones. The before-and-after pair posts to [Slack](/integrations/slack), so anyone who wonders why a number moved can see exactly what changed and when. The paste-and-pray Monday ritual retires, and the leadership meeting starts from numbers that arrived by pipeline, with a visible history, instead of numbers that arrived by keyboard.
The intake sheet as a queue
A Rows table serves as a lightweight work queue: requests appended by one flow, processed by another. The processing agent runs Read Cells over the status column, picks up unprocessed rows, executes the work against [HubSpot](/integrations/hubspot) or the billing system, and writes completion status back to each row's status cell with Write Cells scoped to that single cell, the one case where an overwrite is precisely the point. The spreadsheet is the UI, and the agent is the worker behind it. Anyone on the team can inspect the queue, reprioritize by editing a cell, or add a request by typing a row, no ticket system required.
Human-in-Loop Highlight
Write Cells replaces every value in its target range, and a spreadsheet range is often the most fragile thing in a company's operational stack: the assumptions block a revenue model reads from, the rate table a quote references. There is no merge and no undo through the API; whatever was in A1:D10 is gone the moment the call lands. FlowRunner agents therefore treat overwrites as a two-step act: Read Cells snapshots the range first, and the write proceeds only when the change fits expected bounds. When it does not, the new batch is 30 percent off yesterday's, the shape does not match the range, the target table's ID changed since last run, the agent posts the snapshot and the proposed values side by side and waits for the sheet's owner. Append Rows needs no such ceremony because it destroys nothing. The gate sits exactly where the destruction is.
Agent Capabilities
5 actionsSpreadsheets
2- List Spreadsheets Lists the workspace's spreadsheets with ID, name, and metadata, paginated by cursor. The discovery step for flows that target workbooks by name.
- Get Spreadsheet Retrieves a spreadsheet's pages and tables by ID. The resolution step that turns "the ops tracker" into the concrete table ID and range every read and write needs.
Cells
3- Read Cells Reads a range's values in A1 notation, a block, a column, or a whole area, returned as a two-dimensional array. Feeds downstream steps and snapshots ranges before overwrites.
- Write Cells Overwrites a range with a two-dimensional array of values whose shape matches the range. The deliberate, destructive write this page's gate supervises.
- Append Rows Appends rows after a table's last populated row without overwriting anything. The additive default for logs, syncs, and intake queues.
Frequently Asked Questions
What can FlowRunner do with Rows?
FlowRunner agents can run List Spreadsheets, Get Spreadsheet, and Read Cells in Rows, plus 2 more actions.
Does connecting Rows to FlowRunner require OAuth?
No. Rows connects to FlowRunner with an API key, no OAuth flow required.
Can Rows trigger a FlowRunner workflow automatically?
Rows doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Rows
$100 in credits. No card required. Connect in minutes.