Narrative disclosures

The figures on a financial statement are only part of what a filer tags. The rest is narrative: accounting policies, commitments, segment tables, the notes that explain how a number was arrived at. Arche extracts these as individual disclosures, each tied to the filing it came from.

What a disclosure is

One XBRL text block from one filing. The concept names which block it is, using the filer's own taxonomy, and char_count says how large the text is without making you fetch it.

List a company

curl -X GET "https://api.arche.fi/v1/edgar/companies/0000789019/disclosures?page_size=50" \
  -H "X-Api-Key: $ARCHE_API_KEY" \
  -H "X-Request-ID: $(uuidgen)" \
  -H "Accept: application/json"

Representative response:

{
  "page": 1,
  "page_size": 50,
  "total": 3457,
  "items": [
    {
      "disclosure_id": "e5107ac3-978a-4630-84d9-1f7e569f9722",
      "accession_id": "0001193125-26-027207",
      "cik": "0000789019",
      "period_of_report": "2025-12-31",
      "concept": "src:ContractWithCustomerLiabilityBySegmentTableTextBlock",
      "title": "Contract With Customer Liability By Segment Table",
      "char_count": 235
    }
  ]
}

Narrow with period_of_report for one reporting period, or accession_id for one filing.

Read one disclosure

curl -X GET "https://api.arche.fi/v1/edgar/disclosures/e5107ac3-978a-4630-84d9-1f7e569f9722" \
  -H "X-Api-Key: $ARCHE_API_KEY" \
  -H "X-Request-ID: $(uuidgen)" \
  -H "Accept: application/json"

The detail response adds two fields to the listing shape:

Name
disclosure_text
Type
string
Description
The narrative text itself. Note the field name — it is disclosure_text, not text.
Name
extracted_at
Type
timestamp
Description
When Arche extracted the block from the filing.

Listings omit the text

A listing returns metadata only. This is deliberate: one filing can carry hundreds of kilobytes across its text blocks, and a page of 50 disclosures with full text would be a very large response for a browsing request.

char_count is there so you can decide what is worth fetching before you fetch it. Filter on it to skip the boilerplate and pull the substantial notes.

Fetch disclosure text one record at a time, and let char_count drive the decision. Looping the detail route across a whole listing is the most reliable way to exhaust a read budget on this API.

Finding what you want

concept is the filer's own tag, so it is the precise handle but not a consistent one across companies — different filers use different taxonomies, and a src: prefix means a filing-specific extension rather than a standard us-gaap: concept. title is the human-readable rendering of the same thing.

Match on concept when you are working within one filer's history, and expect to match on title or text when you are comparing across filers.

Next steps

Company identity

Resolve tickers to full issuer records, search by name, and read mapping validity.

Time series

Pull a metric across a whole reporting history, point-in-time or as-reported.

Metric catalog

Look up canonical and derived metric names, formulas, bundles and segments.