Grist

Connect AI agents to Grist, the relational spreadsheet. Agents run record CRUD including upsert, manage tables and columns, discover documents, and run read-only SQL queries.

15 actions available

A source dataset changes in another system on a schedule

Agent reads the source rows and maps them to Grist column ids

Agent runs Add or Update Records to upsert each row keyed on a unique column

Agent runs Query with SQL to aggregate the updated data

Agent reads the aggregate for the report

Agent posts the summary to the team via Slack

Any Delete Records above a threshold pauses for approval

What This Integration Enables

Agents keep a Grist table in sync with an external source using Add or Update Records, log or archive events by appending rows, report on document data with read-only SQL, and provision tables and columns before writing data. The upsert takes items carrying a `require` set (the columns that identify a record) and a `fields` set (the values to write), with optional add-only or update-only toggles, which makes it well suited to idempotent syncs. Records are keyed by column id, and a table's id is also its display name in Grist.

Without FlowRunner

Two systems drift apart A Grist table and its source fall out of sync between manual updates

Reports pieced together Someone joins tables by hand to build a summary

Bulk deletes unguarded A batch record delete has no checkpoint

With FlowRunner

Idempotent upserts Add or Update Records keeps Grist in sync keyed on a unique column

SQL over the document Query with SQL joins and aggregates across tables in one step

Guarded deletes Delete Records above a threshold pauses for human approval

Use Case Scenarios

Keeping two systems in sync

The agent pulls a source dataset from another connector and calls Add or Update Records to upsert each row into a Grist table keyed on a unique id such as an email. Rows that exist are updated; new ones are created. The two systems stay in step without a manual reconcile.

Relational reporting

The agent calls Query with SQL to aggregate document data, joining across tables with standard SQLite syntax, then posts the summary to the team channel with Slack. The report that used to be a manual join now runs as a flow step, and only SELECT is permitted so the query cannot mutate the document.

Cleanup with a human gate

An agent is asked to remove a set of rows. Before it runs Delete Records above the configured threshold, it does not act on its own. It routes the row ids and a count for approval and deletes only after a person confirms.

Human-in-Loop Highlight

Grist upserts and appends are safe to repeat; a bulk Delete Records is the operation that removes data a team may still depend on. FlowRunner's answer is human-in-the-loop, an execution pattern where the agent pauses on its own, assembles the context and the choices, routes to a human on their preferred channel, and resumes the moment they respond. When a Delete Records would remove more rows than the configured threshold, the agent pauses before executing and asks through Slack: "This Delete Records on the `Contacts` table would remove 210 rows. Here is a sample. Approve, narrow the set, or cancel?" The delete runs only after a person confirms, with the approver and timestamp captured in the run log. A connector can delete any records; an orchestration layer knows which deletes should stop and ask.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via Slack

Human decides

Agent resumes with decision

Agent Capabilities

15 actions

Records

Add Records Appends new rows to a table, for logging or archiving incoming events.
Add or Update Records Upserts rows keyed on identifying columns: matching rows are updated, others are created. Optional add-only or update-only toggles make it well suited to idempotent syncs.
Delete Records Deletes rows by their numeric row id.
List Records Reads rows from a table, keyed by column id.
Update Records Updates rows by their numeric row id.

Tables

Create Table Creates a table in a document, for provisioning structure on the fly.
List Tables Lists the tables in a document.
Modify Table Modifies a table's definition.

Columns

Add Columns Adds columns to a table.
List Columns Lists the columns of a table with their ids.

Documents & Workspaces

Get Document Returns the metadata of one document.
List Documents Lists the documents in a workspace.
List Workspaces Lists the workspaces available to the API key, for discovery.

SQL

Query with SQL Runs a read-only SELECT over the document, joining and aggregating across tables with SQLite syntax. Only SELECT is permitted.

Attachments

List Attachments Returns attachment metadata for a document. Uploading attachments is not supported.

Start building with Grist

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us