SEC Filings

Turn SEC filings into structured data

Name the companies, periods and metrics. sieve reads the 10-K, 10-Q or 8-K, pulls the numbers into your schema, and delivers a table through the API, a spreadsheet, or email.

Q1Q2Q3Q4
Company Name
Total Revenue
Revenue Chart
Net Income
Fiscal Quarter

Extracted Data

Company Name

Acme Corporation

Total Revenue

$394.3B

Revenue Chart

Quarterly Revenue Data:

Q1$88.2B
Q2$97.3B
Q3$89.5B
Q4$119.3B

Net Income

$96.99B

Fiscal Quarter

Q4 2023

Extraction Complete5/5 fields

Send us a filing and the fields you need

Tell us the companies, forms and metrics. We will run them and send you the table, usually overnight.

Our clients

Trusted by leading hedge funds

Powering data-driven decisions for the world's most sophisticated investment firms

$100B+

in Assets Under Management

100K+

Data Points Extracted & Validated

XBRL covers the statements. Your model needs the rest.

The SEC XBRL APIs are free and work well for tagged line items. The numbers research teams argue about usually sit somewhere else: segment tables, footnotes, non-GAAP reconciliations, operating KPIs in the MD&A, and press-release exhibits. Paid feeds that cover them arrive late or wrong, and someone on the team checks them by hand every morning.

  • •Untagged KPIs: backlog, same-store sales, units shipped and hashrate are reported in text and tables, not in XBRL.
  • •Footnotes and segments: the detail lives in tables that change shape from one filing to the next.
  • •Vendor errors: one missing or mistyped value in a paid feed starts a correction cycle measured in days.
  • •Small caps and foreign filers: companies outside the big terminals' coverage, with filings only as HTML or PDF.

How it works

1

Tell us the coverage

Tickers or CIKs, form types, periods, and the metrics you want, in plain language. Add the definitions your analysts already use.

Start with one company and one metric.

2

sieve reads the filing

The agent reads the tables and the text, then maps every value to your field names and units. A layout change between filings does not need a new template.

Tables, footnotes, exhibits.

3

You get a table back

CSV or JSON through the API, a spreadsheet, or email. Save the run as a monitor and it re-runs daily, so new filings are picked up without anyone asking.

Daily monitors, email or webhook.

Your schema, not a vendor taxonomy

You name the fields and the units. Nothing to map from someone else's data model onto yours.

Beyond the tagged line items

Segment tables, footnotes, MD&A KPIs and exhibits come back as rows and columns, next to the statement data.

Checked before it reaches your model

Automated checks run on every extraction. On enterprise plans, reviewers compare each value with the filing before it is delivered.

Called from your own code

Ticker, period and metrics in; values out. It drops into the code path that currently emails an analyst when a check fails.

What teams pull from filings

  • Revenue and operating income by segment from 10-K and 10-Q notes
  • Operating KPIs reported only in the MD&A or the earnings release exhibit (EX-99.1)
  • ESG metrics from annual reports and other financial filings
  • Executive compensation tables from DEF 14A proxy statements
  • Figures from small-cap and non-US issuers outside the major data vendors' coverage

One request, one table

The same extraction through the API. Point the agent at a filing, list the columns, and poll for the CSV. Setup and full examples

start = requests.post(f"{BASE_URL}/api/scrapes", headers=headers, json={
    "instruction": "Extract net sales and operating income by reportable segment for each fiscal year shown",
    "target_urls": ["https://www.sec.gov/Archives/edgar/data/320193/000032019324000123/aapl-20240928.htm"],
    "fields": ["segment", "fiscal_year", "net_sales_usd_m", "operating_income_usd_m"],
})
session_id = start.json()["session_id"]
# Poll GET /api/scrapes/{session_id} until status == "done", then download the CSV.

Our investors

Backed by the best

backed by world-class investors who believe in our vision

Y Combinator logo

Questions people ask

Use it for what it covers. The XBRL company-facts API returns tagged financial statement line items for free. sieve is for the values that are not tagged or are tagged inconsistently: segment and footnote tables, KPIs in the MD&A, press-release exhibits, and filers with sparse tagging.

10-K, 10-Q, 8-K and their exhibits, 20-F, 40-F, S-1, DEF 14A proxies and 13F, plus annual reports and investor documents published outside EDGAR. If it is a document or a web page, we can extract from it.

Automated checks run on every extraction and report coverage gaps and outliers on the run. On enterprise plans, reviewers compare each value with the source filing before delivery, which is how the quant team in our case study stopped checking its feed by hand.

Yes. Save a run as a monitor and sieve re-runs it every day, then sends the new values by email or webhook. A filing posted today is in your data tomorrow.

2025 Sieve Data Inc. All Rights Reserved.