Skip to main content
The Exports API gives you paginated, read-only access to the ClearGrid datasets that belong to your lender. Every row returned is scoped to the calling lender and passed through an allowlist that strips and masks sensitive fields, so you can pull operational data into your own warehouse without handling raw PII. There are two endpoints. The catalog endpoint lists what is available to you, and the fetch-dataset endpoint retrieves the rows. Both endpoints use the same cursor-based pagination, which supports an initial full bootstrap and ongoing incremental delta loads from one consistent interface.
The catalog lists 25 dataset codes across five sections. As of v1.2, 19 codes are available to fetch today. The remaining 6 codes (D04, D06, D10, D11, D13, D24) are reserved for future use and listed with available: false. Fetching any dataset with available: false returns a 404.

​
Endpoints

Catalog endpoint

Discover the sections and datasets you are allowed to export, with cursor fields and availability flags.

Fetch dataset endpoint

Retrieve paginated, redacted rows for a single dataset, for both initial and incremental loads.

​
Available datasets at a glance

​
Date and time formats

Exports return two distinct date conventions, depending on how the field is stored at source — parse them with separate handling:
  • System timestamps — created_at, updated_at, communicated_at, and any genuine Date-typed field (e.g. the PTP / broken-PTP dates on D16) are ISO-8601 with milliseconds and a Z suffix (e.g. 2026-05-28T11:42:08.000Z).
  • Business date columns — certain date fields are stored as plain strings and are returned exactly as stored, in DD/MM/YYYY (e.g. 22/02/2026). They are not normalized to ISO-8601, carry no time component, and no timezone. Confirmed DD/MM/YYYY fields today: D02 / D03 submitted_date and due_date, and D16 dpd_due_date and last_due_date.
If you ingest into a typed warehouse (e.g. BigQuery), do not map the business date columns to a TIMESTAMP/DATE type expecting ISO-8601 — ingest them as STRING and parse DD/MM/YYYY, or transform on the way in. Mapping them as native dates will fail or silently misparse (e.g. day/month inversion).

​
Quick reference

Pagination

Cursor encoding, initial vs. incremental loads, and variable page-size behavior.

Errors

Every status code the export endpoints return, with example bodies.

Datasets

Field-by-field reference for each Phase 1 dataset.

Need Help?

Our support team is available Monday-Friday, 9 AM - 6 PM GST. Reach out at support@cleargrid.co or visit our help center.