Skip to main content
GET
One page of an exposed dataset, exposed columns only

Authorizations

Authorization
string
header
required

A project Data API key (hdk_…), created in the project's Settings → Data API.

Path Parameters

dataset
string
required

Query Parameters

limit
integer
default:25
Required range: 1 <= x <= 1000
offset
integer
default:0
Required range: x >= 0
cursor
string

Opaque next_cursor from a previous cursor-mode page (or X-Next-Cursor from CSV). Cursor pages order by received_at DESC, rowid ASC, using the records time index's implicit rowid tie-break.

pagination
enum<string>

Set to cursor on the first page to opt into keyset pagination and bounded totals; continuation requests use cursor and remain bounded.

Available options:
cursor
since
string<date-time>

Only records received at or after this ISO 8601 time.

until
string<date-time>

Only records received before this ISO 8601 time.

format
enum<string>

csv for a UTF-8 CSV (with BOM, formula-safe) instead of JSON; X-Total-Count is exact for legacy offset requests and capped at 10,000 in cursor mode, X-Total-Count-Truncated marks the cap, and cursor mode returns X-Next-Cursor.

Available options:
json,
csv

Response

Legacy offset pages are newest first by received_at DESC, id DESC and return exact totals. Cursor pages use received_at DESC, rowid ASC to seek in the existing time index; set pagination=cursor on the first request. Cursor-mode totals are capped at 10,000 with total_truncated=true when larger.

dataset
string
required
columns
string[]
required
rows
object[]
required
total
integer
required

Exact for requests without cursor mode. In cursor mode, capped at 10,000.

total_truncated
boolean
required

True only in cursor mode when the dataset has more than 10,000 matching rows.

limit
integer
required
offset
integer
required
next_cursor
string | null
required

Cursor-mode continuation position, or null for legacy offset requests and the last cursor page.