SEC Filing Search — Full-Text EDGAR Filings Scraper

Search the full text of every SEC EDGAR filing since 2001 and export the matches as structured data — company, tickers, CIK, form type, filing and period dates, 8-K items, SIC code, state of incorporation and a direct link to the document itself.

This is the search behind EDGAR's full-text interface, turned into a dataset. Ask "which companies disclosed a material weakness in their last 10-K?" and get a table instead of a web page.

No API key required.

What you get

Field	Description
company	Filer name, parsed clean
tickers	Ticker symbols, e.g. ["FITB","FITBI"]
cik, cikPadded	SEC company ID, both trimmed and zero-padded
allFilers	Every filer on a multi-filer submission
form, rootForm	Exact form and its base type, e.g. 10-K/A → 10-K
fileType, fileDescription	Which document inside the filing matched
filedAt	Filing date
periodEnding	Fiscal period the filing covers
items	8-K item codes, e.g. ["5.02"] for an executive change
sic	Industry classification code
businessState, businessLocation, incorporatedState	Where the company operates and is incorporated
accessionNumber, fileNumber	SEC identifiers
relevanceScore	EDGAR's own match score
documentUrl	Direct link to the matching document
filingIndexUrl	The full filing's index page

Example input

{
  "query": "material weakness",
  "exactPhrase": true,
  "forms": "10-K",
  "startDate": "2026-01-01",
  "maxFilings": 200
}

Monitor a theme across 8-Ks:

{
  "query": "artificial intelligence",
  "forms": "8-K",
  "startDate": "2026-01-01",
  "endDate": "2026-06-30"
}

Watch specific companies:

{ "query": "going concern", "ciks": ["320193", "1318605"] }

Example output

{
  "company": "FIFTH THIRD BANCORP",
  "tickers": ["FITB", "FITBI", "FITBO", "FITBP"],
  "cik": "35527",
  "form": "10-K",
  "filedAt": "2022-02-25",
  "periodEnding": "2021-12-31",
  "businessState": "OH",
  "businessLocation": "Cincinnati, OH",
  "sic": "6022",
  "accessionNumber": "0000035527-22-000119",
  "documentUrl": "https://www.sec.gov/Archives/edgar/data/35527/000003552722000119/fitbannualreport202110-k.pdf"
}

Who uses this

• Investment research — find every company that used a phrase before it became consensus: "supply chain constraints", "AI capex", "stablecoin"
• Short sellers and credit analysts — going concern, material weakness, restatement and covenant breach as a screen, not a hunch
• Corporate lawyers — precedent language across real filings, and who filed it
• Risk and compliance — monitor a competitor's or supplier's 8-K items as they are filed
• Academics and data teams — build a labelled filing corpus with document URLs to fetch
• M&A and BD teams — 8-K item codes surface executive changes, acquisitions and material agreements the day they publish

A common setup is a daily run with startDate set to yesterday and a watchlist phrase, feeding a webhook.

Two limits, stated plainly

• Full-text search starts in 2001. Older filings exist in EDGAR but are not in the full-text index, so a phrase search cannot reach them. For pre-2001 work you need the filing archives directly, not this search.
• EDGAR serves at most 10,000 results per query. It is an OpenSearch index with a hard result window. When your query matches more, the actor reports the true total and tells you it can only retrieve the first 10,000 — the fix is to narrow by form type or split the date range into chunks and run each.

There is a third quirk worth knowing if you ever call this API yourself: when you exceed the result window, EDGAR replies with HTTP 200 and an error body rather than an error status. A naive client reads that as success and silently returns nothing. This actor detects it and says what happened.

The contact e-mail field

The SEC requires automated clients to identify themselves with a contact address and throttles or blocks generic agents. The actor always sends one, but putting your own e-mail in the Contact e-mail field identifies your runs properly and makes them more reliable — recommended for anything beyond a quick test.

Pricing

Pay per result. Each matching filing document counts as one result.

Notes

• One filing can match on several documents; each matching document is its own row, identified by fileName, with accessionNumber shared. Deduplicate on accessionNumber if you want one row per filing.
• exactPhrase wraps your query in quotes, which is how EDGAR distinguishes a phrase from loose terms. Turn it off for broader recall.
• This is an unofficial actor and is not affiliated with the SEC.