Get Document Page

Return one document page's OCR text in reading order and, optionally, its layout regions. Addressed by the same `document_id` returned by [List Documents](/reference/v3/documents) and [Query](/reference/v3/query), plus a `page_number`. No new ID scheme. **v1 is OCR-paged documents only.** A document with no OCR page data (for example, a `.txt` or `.docx` file) returns a `404` with code `document_has_no_pages`; use [Get Document](/reference/v3/document) or [Get Chunk](/reference/v3/documents/chunk) for those. A blank in-range page returns `200` with empty `text` and a `page_empty` warning, not a `404`. A document still indexing with partial OCR present returns `200` with a `document_indexing` warning rather than failing.

Path parameters

collection_namestringRequired
document_idstringRequired
page_numberintegerRequired>=0

Query parameters

include_regionsbooleanOptionalDefaults to false

Response

Get Document Page response.
collection_namestring
document_idstring
pageinteger
The requested page number, echoed back.
textstring

The page’s full text in reading order. Empty string when the page exists but has no extractable text (see warnings).

page_countinteger or nullOptional

Best-effort total page count, derived from the document’s OCR data. Null when it cannot be determined.

text_sourceenumOptionalDefaults to ocr

Where text came from. Always ocr today; a future value would mean reconstructed text, not primary-source OCR.

regionslist of objects or nullOptional

Layout regions on this page when include_regions is true. Regions of type figure, chart, or image carry a presigned image_url.

warningslist of objectsOptional

Non-fatal conditions on this response, such as document_indexing (content may be incomplete) or page_empty (page exists, no extractable text).

© 2026 Captain