Tax return summary document profile
Index filed or prepared tax returns.
The Tax return summary profile helps you index filed or prepared tax returns. It belongs to the personal finance family.
Recognize the source
A return or draft return, not tax advice or a tax assessment; record exact form and status. Do not infer eligibility or liability.
Review the document itself before applying the profile. Similar labels, dates, parties, or amounts can appear in another document type with a different meaning. Use the source context to distinguish them.
Expected fields
Core fields form the profile’s main structure. Recommended and optional fields add supported detail. None of these designations permits guessing a value that is absent or ambiguous.
| Field | Type | Inclusion | Meaning and extraction rule |
|---|---|---|---|
| Taxpayer reference | text | Core | The source-stated taxpayer reference for this document and its identified subject. Copy exact source spans for this one scalar field, preserving their order, ownership and qualifications. Ordered spans may continue across segments. Do not synthesize a summary or combine unrelated fields or repeated records; return unresolved when the scalar meaning is ambiguous. |
| Tax jurisdiction | text | Core | The source-stated tax jurisdiction for this document and its identified subject. Copy exact source spans for this one scalar field, preserving their order, ownership and qualifications. Ordered spans may continue across segments. Do not synthesize a summary or combine unrelated fields or repeated records; return unresolved when the scalar meaning is ambiguous. |
| Tax year | text | Core | The source-stated tax year for this document and its identified subject. Copy exact source spans for this one scalar field, preserving their order, ownership and qualifications. Ordered spans may continue across segments. Do not synthesize a summary or combine unrelated fields or repeated records; return unresolved when the scalar meaning is ambiguous. |
| Form name | text | Core | The source-stated form name for this document and its identified subject. Copy exact source spans for this one scalar field, preserving their order, ownership and qualifications. Ordered spans may continue across segments. Do not synthesize a summary or combine unrelated fields or repeated records; return unresolved when the scalar meaning is ambiguous. |
| Return submission status | text | Core | The explicitly stated submission state of this return, such as draft, filed or accepted. This is not a tax filing category such as single or joint. Do not infer submission from the existence or signature of a form. |
| Stated tax due | number | Optional | The source-stated tax due for this document and its identified subject. Use an explicitly stated value with safe numeric normalization only. Preserve its currency, unit, scale and basis in evidence metadata; do not calculate, estimate or combine values. |
| Stated refund | number | Optional | The source-stated refund for this document and its identified subject. Use an explicitly stated value with safe numeric normalization only. Preserve its currency, unit, scale and basis in evidence metadata; do not calculate, estimate or combine values. |
| Currency | text | Optional | The currency explicitly associated with the selected monetary fields. Preserve source representation; do not infer a currency from an ambiguous symbol or company location. Conflicting currencies require separate fields or unresolved values. |
Missing, ambiguous, and repeated values
Missing or ambiguous fields remain null with a reason. Null is not zero, false, an empty statement, or proof that a document-wide fact does not exist. Inspect unresolved states and their supporting passages before accepting the table. This document type can contain repeated records such as schedules and tax line items. The profile uses one row per document; do not compress repeated records into a scalar field or confuse a line value with a document-level total. Additional supported fields can be considered when the extraction experience allows them. Review their definitions and types explicitly; changing a field’s meaning requires a new extraction preview.
Review the result
Select evidence-backed cells and read the retained source. Text should preserve the relevant source spans and qualifications. Numbers and dates must be stated unambiguously and normalized safely; do not calculate unstated values or infer missing units, currency, scale, date parts, or time zones.
Verify the result against the captured document before acceptance. Source support does not independently establish perfect OCR, exhaustive field discovery, or correct domain interpretation.
See profile modes, table review, and captured evidence.