Tables

A table is data your flows read and write between runs: posts already answered, products found, daily counts, leads from a form. There are two kinds:

Rows: scope, key, label

Every row has, besides its columns:

Scope

A table's scope decides which rows a run sees:

Open a table to see each row's scope, key and label. Clear test data removes what test runs wrote.

The Tables page

Dataset › Tables has two tabs (?tab=), and its address holds the state (?page, ?size, ?q, ?sort, ?dir), so a list can be shared and Back works:

Open a table for its rows (?show=all\|real\|test), a page at a time. On a table of yours you can add and edit rows; on a node's own you can browse and delete.

Node tables

A node definition declares the tables it owns (table in ); they live in code (lib/flow/table, one registry for every node), not in the database, are named <node type>/<name> and keep their definition's columns (you can browse and delete rows, not change columns). Their rows are kept in the database. Built-in nodes reach them through one helper, flowTableCall (lib/flow/table/call), bundled into the node. Built in:

Create your own table

Dataset › Tables › New table: a Name and its Columns, each with a type:

Open a table to browse, add, edit and delete rows by hand. The names id, key, label and campaignId are row fields, not columns.

Use it in a flow: the Data table node

Examples:

{ "operation": "find", "where": { "key": "{{ nodes.loop.output.item.id }}" }, "limit": 1 }

{ "operation": "insert", "data": { "key": "{{ nodes.loop.output.item.id }}", "label": "seen", "url": "{{ nodes.loop.output.item.url }}" } }

{ "operation": "upsert", "where": { "key": "{{ nodes.input.output.keyword }}" }, "data": { "link": "{{ nodes.affiliate.output.link }}" } }

Node output (row is the first row, for convenience):

{ "rows": [{ "id": "6ae1…", "key": "2115091199661146286", "label": "seen", "url": "https://x.com/…", "createdAt": "2026-10-28T06:11:18.000Z" }], "row": { "id": "6ae1…", "key": "2115091199661146286", "label": "seen", "url": "https://x.com/…", "createdAt": "2026-10-28T06:11:18.000Z" }, "count": 1 }

A find with no rows leaves through the empty dot; everything else through next.

Test runs write too, as test rows. They read and write their own rows, never the real ones; clear them with Clear test data.

From a script node

await fetch(`${api.link}/api/run/table`, {
	method: "POST",
	headers: { authorization: `Bearer ${api.token}`, "Content-Type": "application/json" },
	body: JSON.stringify({ table: "6ae1…", operation: "find", where: { key: "123" }, limit: 1, sort: { column: "repliedAt", direction: "desc" } }),
});

A node's own table by key, and a claim (insert only keys not taken in the scope, held for 15 minutes):

body: JSON.stringify({ table: "x/collect/post", operation: "claim", data: [{ key: "2115…", label: "claimed", url: "…" }], holdMs: 900_000 })

See the .

FAQ

What happens to flows when I delete a table?

The table and all its rows are removed, and Data table nodes using it fail until you pick another. A node's own table can't be deleted: delete its rows, or Clear test data.

Can I change columns later?

Yours, yes, with Edit columns. Existing rows keep their values; a removed column stops being read and written. A node's own table has its columns from its definition.

Why don't my test runs see real rows?

On purpose: they read and write test rows, so a test can't skip a post a real run took, nor block one.