<!-- Agent Datasets docs · stocks_get_filing_section · canonical: https://www.agentdatasets.com/docs/tools/stocks_get_filing_section · rendered from https://www.agentdatasets.com -->

# stocks_get_filing_section

Read one 10-K/10-Q section in bounded, lossless text pages.

| Field | Value |
| --- | --- |
| Category | Stocks (`stocks`) |
| Exposure | public |
| MCP tool | `stocks_get_filing_section` |
| Documentation | https://www.agentdatasets.com/docs/tools/stocks_get_filing_section |

## Description

Read one 10-K/10-Q section in bounded, lossless text pages.

Use this when you know the filer and need the actual text of one filing
item, such as Item 1A risk factors or Item 7 MD&A. Use
`stocks_get_sec_filings` for filing metadata and sec.gov links without
document text. Use `stocks_search_filings` for cross-filer full-text
discovery when you do not yet know the accession and item to read.

The shared envelope's `data` identifies the company and filing, echoes
the canonical `item_code` plus its common title, and returns `text` with
its exclusive `[char_start, char_end)` range, total
`section_char_count`, and `document_url`. `meta` carries
`source: "sec-edgar-archives"` and the section extraction timestamp.
`pagination.limit` is the requested character cap.

To page, keep the filing selector, item, and `page_chars` unchanged and
pass `pagination.next_cursor` as `cursor`. Concatenating each page's
`data.text` until `has_more` is false reproduces the stored section
exactly; the cursor is opaque and only moves forward.

If the section has not been extracted, this call synchronously attempts
the existing on-demand SEC extraction path. Errors carry a machine code:
`unknown_entity` for an unknown ticker/CIK, `bad_parameter` for invalid
addressing, item codes, bounds, mismatched accession/company pairs, or
cursors, and `out_of_coverage` for unindexed/unsupported filings or an
extraction that cannot provide the requested item. Coverage errors link
to sec.gov so the source can be read directly; text is never fabricated.

Example: `stocks_get_filing_section(ticker_or_cik="AAPL",
form_type="10-K", item="Item 7", page_chars=10000)`.

## Parameters

Input schema advertised to MCP clients and mirrored by the REST query string.

| Parameter | In | Type | Required | Default | Constraints | Description |
| --- | --- | --- | --- | --- | --- | --- |
| ticker_or_cik | input | `string` | yes | — | — | A US public company by exact ticker ('AAPL', 'BRK.B') or SEC CIK ('320193' or zero-padded). The filing must belong to this company. |
| item | input | `string` | yes | — | — | Filing item to read. Canonical codes include ITEM_1A and PART_I_ITEM_2; friendly forms such as '1A', 'Item 7', and 'Part I Item 2' are normalized. |
| accession_number | input | `string \| null` | no | null | — | Exact SEC accession (NNNNNNNNNN-NN-NNNNNN). Pass this OR form_type, never both; use it for a specific filing. |
| form_type | input | `string \| null` | no | null | — | Select the latest indexed filing of this exact form. Supported values are 10-K and 10-Q. Pass this OR accession_number, never both. |
| page_chars | input | `integer` | no | 10000 | min 1000, max 25000 | Maximum characters in this page (1000–25000); paragraph boundaries may make a page shorter. Keep it constant while paging. |
| cursor | input | `string \| null` | no | null | — | Opaque forward-only cursor from pagination.next_cursor. Pass it back unchanged with the same filing, item, and page_chars; omit it for the first character page. |

## Provenance

- `sec-edgar-archives` — U.S. Securities and Exchange Commission (EDGAR), Section text extracted from 10-K/10-Q primary documents on sec.gov Archives (canonical Items with accession-number and source-URL lineage); full document HTML is never stored (U.S. Government work, public domain (17 U.S.C. § 105)). See https://www.agentdatasets.com/docs/attribution.md

Dataset registry: `stocks_filing_sections`.
