SEC Standardized Text

Requires a SEC Standardized Text subscription.
Every SEC html, (most) pdf, and text files converted to columnar format.
Datasets and SQL queries are updated as new filings come up.

Daily datasets are combined at 06:00 UTC Tuesday–Saturday. For daily downloads, we recommend pulling after 09:00 UTC.

Data

accessionNumber + sequence identifies a document. id identifies the order of elements within a document.

Table (16 rows)

Column	Meaning	Example
accessionNumber	Filing accession number`	`110465923015159`
sequence	Document order in the filing	`23`
filingDate	filing date	`2026-01-02`
filename	filename	`filename2.txt`
documentType	SEC document type	`LETTER`
id	Sequential output row number, starting at `0`	`3`
level	Nesting depth within the document	`2`
type	`title`, `text`, `textsmall`, `table`, `image`, `list`, or `redundant`	`"table"`
text	Text content; image alt text for images; associated title for tables. `null` for lists	`"Consolidated Balance Sheets"`
matrix	Tables: rows of cell strings. Lists: `[relative_depth, item_text]` pairs, both strings. `null` otherwise	`[["Year", "Revenue"], ["2025", "100"]]`
cleaned	Table cleanup changes and unresolved issues, represented as a list of strings. `null` for other types	`["removed_empty_rows", "merged_header_rows"]`
preamble	Table rows only: associated text before the table, or `null`	`"Amounts in millions."`
footnotes	Table rows only: list of `[id, text]` pairs, or `null`	`[["(1)", "Includes discontinued operations."]]`
postamble	Table rows only: associated text following the table’s footnotes, or `null`	`"See accompanying notes to the financial statements."`
class	Classification assigned by the SEC rules	`"item"`
standardizedTitle	Normalized identifier for a recognized heading; otherwise `null`	`"item1a"`

Download Dataset

Download a filing year, including newly processed filings:

from datamulehub import datasets

datasets.download(
    "standardized_text",
    year=2026,
    filename="standardized_text_2026.parquet",
)

Use from_storage="daily" for the published yearly snapshot, or "delta" for recent per-filing outputs only.

Download all parsed documents from one filing:

datasets.download(
    "standardized_text",
    accession=["0000320193-24-000123"],
    filename="filing_text.parquet",
)

Query Database

Athena exposes column names in lowercase: standardizedTitle in Parquet is standardizedtitle in SQL.

Find risk-factor headings in 10-K documents:

from datamulehub import databases

rows = databases.read_query(
    """
    SELECT *
    FROM standardized_text
    WHERE documenttype = '10-K'
      AND type = 'table'
      AND filingdate >= '2024-01-01'
      AND filingdate < '2025-01-01'
    LIMIT 10
    """
)

print(rows)

Use databases.query(sql, output_dir="results") to save query results as Parquet files instead.