API · Filings

Resolve a filing page snapshot

Turn a page_id into a citable URL — rendered PNG, EDGAR link, or ESEF viewer link.

GET/api/v1/data/filings/{page_id}/snapshot

page_ids are returned by other endpoints alongside any figure or passage that has a filing-level source. Pass that id here and you get back a viewable URL plus the surrounding metadata (company_name, year, document_type, page_number).

See Page IDs and citations for the full lifecycle including stable-CDN vs. presigned-S3 URLs and caching guidance.

When to use this

Whenever you render a number or quote to a user. A clickable "see source" link transforms a generic financial product into one auditors and CFOs trust.

Path

FieldTypeDescription
page_idrequiredstringPage identifier of the form <document id>_<page>, as returned by content search, statement cells, financials provenance, or Nordic revenue geography rows.

Query

FieldTypeDescription
dpiinteger
default: 150
Render resolution for Nordic PNG snapshots (72–300).

Example

bash
curl "https://api.clarifo.com/api/v1/data/filings/722df34e9f191a97518a8560755cb9e9_20/snapshot?dpi=200" \
  -H "Authorization: Bearer $CLARIFO_API_KEY"

Response shapes

The kind field determines what snapshot_url points to:

KindDescription
rendered_pngPer-page PNG render of the original PDF filing.
snapshot_endpointNordic filing page without a stored render: snapshot_url is Clarifo's snapshot endpoint, which renders the page on request.
edgar_linkSEC EDGAR link to the filing document for US filings. Also carries the cited passage and a highlight_url that opens the filing at that passage (see below).
viewer_linkLink to the source document (for example an ESEF viewer) for a Nordic filing that cannot be rendered per page.
Nordic — rendered_png
{
"page_id": "722df34e9f191a97518a8560755cb9e9_20",
"available": true,
"kind": "rendered_png",
"snapshot_url": "https://cdn.clarifo.com/snapshots/…/p20.png",
"source_url": "https://…/api/pdf-snapshot?page_id=722df34e9f191a97518a8560755cb9e9_20",
"company_name": "Neste Oyj",
"year": "2024",
"document_type": "annual_report",
"page_number": 20,
"source_kind": "s3_pdf",
"page_number_kind": "pdf_render",
"page_reference": "p. 20",
"page_reference_type": "page",
"dpi": 200,
"citation_label": "Neste Oyj 2024, p. 20",
"citation_markdown": "[Neste Oyj 2024, p. 20](https://cdn.clarifo.com/snapshots/…/p20.png)"
}
US — edgar_link
{
"page_id": "0001045810-26-000075_12",
"available": true,
"kind": "edgar_link",
"snapshot_url": "https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm#part-i-item-2-mda",
"source_url": "https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm#part-i-item-2-mda",
"company_name": "NVIDIA Corporation",
"official_name": "NVIDIA Corporation",
"year": "2026",
"quarter": 2,
"document_type": "10-Q",
"section": "Part I Item 2 - MD&A",
"page_number": 28,
"page_number_kind": "printed",
"source_kind": "sec_edgar",
"page_reference": "Part I Item 2 - MD&A, p. 28",
"page_reference_type": "page",
"excerpt": "27

* * *

Revenue was $96.2 billion, up 106% from a year ago and up 18% sequentially. …",
"highlight_url": "https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm#part-i-item-2-mda:~:text=Revenue%20was%20%2496.2%20billion%2C%20up%20106%25%20from%20a",
"citation_label": "NVIDIA Corporation Q2 2026 10-Q — Part I Item 2 - MD&A, p. 28",
"citation_markdown": "[NVIDIA Corporation Q2 2026 10-Q — Part I Item 2 - MD&A, p. 28](https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm#part-i-item-2-mda:~:text=Revenue%20was%20%2496.2%20billion%2C%20up%20106%25%20from%20a)"
}

US filings: sending a reader to the passage

For US pages the response carries the same evidence fields as content search for the same page_id:

  • section and page_number, with page_number_kind: "printed". The number is the page printed in the filing, recovered from the page-break markers in the indexed text. It is omitted when the passage has no marker, rather than guessed; the page_number of 1 that older responses carried was a placeholder.
  • excerpt: up to 500 characters of the cited passage.
  • highlight_url: the section link plus a #:~:text= fragment built from the first sentence of the excerpt (at most 150 characters, percent-encoded). Chrome, Edge and Safari scroll to and mark the text; a browser that does not support text fragments, or a filing whose text differs, simply opens the section.
  • citation_markdown links to highlight_url when present.

highlight_url is only ever an https://www.sec.gov URL built from the stored filing link and the stored passage. Nothing from the request goes into it, and Clarifo does not contact SEC while answering. If the stored link is on any other host, or no passage is stored, highlight_url and excerpt are absent and snapshot_url is the plain link as before. kind, snapshot_url and every earlier field keep their meaning.

Not available
{
"page_id": "722df34e9f191a97518a8560755cb9e9_20",
"available": false,
"company_name": "Neste Oyj",
"year": "2024",
"document_type": "annual_report",
"page_number": 20,
"source_kind": "s3_pdf",
"reason": "Source page preview is not available for this filing yet",
"page_reference": "p. 20",
"page_reference_type": "page",
"citation_label": "Neste Oyj 2024, p. 20",
"citation_markdown": "Neste Oyj 2024, p. 20"
}

When available is false

reason is one of:

reasonMeaning
Page not foundThe page_id does not match any indexed page. Only page_id, available and reason are returned. Page ids have the form <document id>_<page>; an id without the _<page> suffix, such as the bare SEC accession number on US revenue geography rows, is never found.
Source page preview is not available for this filing yetNordic page found, but no render or source link exists yet. Metadata and citation fields are still returned so you can cite the page in plain text.
SEC EDGAR link not yet ingested for this pageUS page found, but its EDGAR link is not stored yet. Metadata and citation fields are still returned.
Snapshot service unavailableTemporary server-side problem; retry later.

When available is false, citation_markdown is the plain label with no link, so a reader is never sent to a page that has no preview. The miss is an HTTP 200 response, like every other tool-level miss in this API.

Citation fields

citation_label and citation_markdown are pre-formatted strings ready to embed in a UI or document. Use citation_markdown directly when rendering Markdown -- it includes the hyperlink to the source.

page_reference is the anchor on its own: "p. 20" for a paginated Nordic PDF (page_reference_type: "page"), a section heading for SEC filings, which are indexed by section rather than page (page_reference_type: "section"), or the document type when neither is known ("document"). page_number_kind: "pdf_render" marks a Nordic page number as the position in the PDF (cover = 1), which can differ from the number printed on the page. source_url is a stable link to the same source, usable when snapshot_url is a short-lived URL.