Skip to main content
POST
Read grid data

Headers

x-project-id
string<uuid>
required

Project ID for tenant scoping

X-Workspace-Id
string<uuid> | null

Overrides the key's workspace

X-API-Key
string | null

User API key

Body

application/json
view
enum<string>
required

Which grid data slice to return: columns, rows, cell (needs row_id/column_id), cell_run (that cell's producing run - needs row_id/column_id/workflow_run_id), or row_search. Prefer cell/cell_run over rows+page_size=1 for a single-row/cell question - but only for expandable (blob-backed) columns, e.g. workflow/output columns with markdown/json/array/file content; plain scalar columns (text/number/boolean/single_select) already return their full value via rows and have no cell view to expand.

Available options:
columns,
rows,
cell,
cell_run,
row_search
grid_id
string<uuid>
required

Grid to read from

row_id
string<uuid> | null

Row id; required for cell and cell_run views

column_id
string<uuid> | null

Column id; required for cell and cell_run views. For cell_run specifically, must be the WORKFLOW-type column id (not the input/output column) - the one whose AI runs are being traced.

workflow_run_id
string<uuid> | null

Run id; required for the cell_run view

query
string | null

Search text; required for the row_search view. row_search matches only searchable (text-like input) cell values — generated workflow output stored as markdown/JSON/blobs is not indexed and won't match.

page
integer
default:1
Required range: x >= 1
page_size
integer
default:50
Required range: 1 <= x <= 200
column_ids
string<uuid>[] | null

Column id filter for the rows view — return only these columns per row instead of every column. See also field_name/column_type for the columns view.

row_ids
string<uuid>[] | null

Row id filter for the rows view — fetch specific, non-contiguous rows (e.g. 'rows 3, 7, and 9') instead of a page range. Resolve ordinal positions to ids with a prior rows view read, then pass them here.

column_type
enum<string>[] | null

Column-type filter for the columns view. Defaults to INPUT columns; pass ['output'] to see the generated/workflow columns the grid produces for each row.

Available options:
input,
workflow,
output
field_name
string | null

Column filter for the columns view — matches the underlying workflow input schema's field name (not the grid's display column name). Use this to answer 'which column maps to field X' directly instead of listing all columns and matching by eye.

workflow_id
string<uuid> | null

Filter columns by workflow id (for the columns view)

Response

Successful Response