BETA / SOURCE-NATIVE XBRL

Keep the context behind each fact.

Read one XBRL XML instance from a verified filing manifest. Facts retain qualified names, exact source strings, context and unit references, decimals, precision, nil attributes and tuple ancestry. Context/unit XML trees retain explicit and typed dimensions with namespaces.

Resolve the document first

  1. Find the issuer’s actual filing accession using filing history.
  2. List its document manifest with /api/v1/documents.
  3. Select an XBRL instance XML filename, including an SEC-extracted instance where available. Do not send a linkbase, arbitrary URL or inline-XBRL HTML document.
curl 'https://sec-api.dev/api/v1/xbrl-instance?cik=320193&accession=0001140361-26-035325&document=ef20081427_8ka_htm.xml&limit=100&offset=0' \
  -H "Authorization: Bearer $SEC_API_KEY"

Response contract

data contains facts, contexts, units and namespaces. Only contexts and units referenced by the current fact page are returned. Use pagination.nextOffset until null. Namespace URIs define identities; prefixes alone do not. Numeric source strings must not be silently coerced into floating-point values.

Bounds and interpretation

The parser accepts at most 2 MB of XML, 64 nesting levels and 100,000 elements. DTD and entity declarations are rejected. The response obeys a 900 KB budget; overlarge input or output returns a documented error rather than a truncated successful statement. Missing references, malformed XML and unsupported roots are rejected.

This is not a taxonomy validator, calculation validator, standardized financial statement service, inline-XBRL HTML parser or cross-filing segment database. A dimension can change a fact’s meaning; compare the full context, unit, period and taxonomy before comparing values. Source provenance identifies the document and retrieval context.

Company Facts provides a different consolidated source. Observation series selects periods from that consolidated source; it does not replace filing-level contexts.