get_filing_text
Read-only
Text of an issuer's filing of a given form — the most recent, or a HISTORICAL one.
period: select an older filing instead of the latest — a fiscal year ("2021") or a period-end
date ("2023-06-30", the quarter/year the filing covers). Omit for the most recent. This is how you
read a specific year's 10-K or a specific quarter's 10-Q, not just today's.
section: business · risk_factors · legal_proceedings · mdna · financial_statements. STRONGLY
prefer a section — a full 10-K is ~55,000 tokens and will swamp your context.
find: jump straight to a keyword ("restructuring", "Value-at-Risk", "pension", a segment name)
instead of reading from the top — returns the window around the first match and reports
data.match_count. This is how you pull ONE number out of a 400k-char filing in a single call;
combine with section, or use it alone when a section isn't detected. To reach a later match, call
again with offset = data.window_start + data.chars_returned.
offset: start position in the (section or full) body, for paging a long section.
Output is always bounded and sets data.truncated when there is more after this window, with
chars_total showing the full length and data.window_start where this window began.
For an earnings press release / adjusted (non-GAAP) figures, use get_filings(kind="press_releases"),
which serves the EX-99 exhibit text — those numbers are not in the 10-K/10-Q body.Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| find | No | ||
| form | No | 10-K | |
| entity | Yes | ||
| offset | No | ||
| period | No | ||
| section | No | ||
| max_chars | No |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |