SEC EDGAR Scraper — Filings, Full-Text RAG & XBRL Financials

Pricing

Rating

Developer

Actor stats

Categories

Share

Extract everything from SEC EDGAR in one Actor: company filings metadata, EDGAR full-text search, section-parsed RAG-ready full text (Item 1A Risk Factors, Item 7 MD&A, …), and XBRL financial facts — with no API key required.

At a glance

Features

Quick Start

Get Apple's latest annual and quarterly reports as RAG-ready chunks:

{
"mode":"fulltext",
"ticker":"AAPL",
"formTypes":["10-K","10-Q"],
"dateFrom":"2024-01-01",
"chunking":"paragraph",
"maxItems":5
}

Modes

Input

¹ filings, fulltext and facts modes require a company; fulltext can alternatively run from a query.

Output

Example: fulltext item

{
"itemType":"filing-fulltext",
"title":"Apple Inc. — 10-K 2025-10-31",
"company":"Apple Inc.",
"ticker":"AAPL",
"cik":"0000320193",
"formType":"10-K",
"filedAt":"2025-10-31",
"accessionNo":"0000320193-25-000123",
"documentUrl":"https://www.sec.gov/Archives/edgar/data/320193/...",
"sections":[
{"name":"Item 1A — Risk Factors","charCount":38241},
{"name":"Item 7 — Management's Discussion and Analysis","charCount":21077}
],
"chunks":[
{"text":"The Company's business, reputation, results of operations...","section":"Item 1A — Risk Factors","order":12}
],
"textLength":220151
}

Example: facts item

{
"itemType":"fact",
"company":"Apple Inc.",
"taxonomy":"us-gaap",
"tag":"RevenueFromContractWithCustomerExcludingAssessedTax",
"unit":"USD",
"value":111184000000,
"start":"2025-12-28",
"end":"2026-03-28",
"fiscalYear":2026,
"fiscalPeriod":"Q2",
"form":"10-Q"
}

Recipes

1. Build a RAG knowledge base from SEC filings

Pull two years of 10-K/10-Q full text for your watchlist, chunked for embeddings. Each chunk carries its Item section, so you can filter retrieval to Risk Factors or MD&A only.

{
"mode":"fulltext",
"ticker":"NVDA",
"formTypes":["10-K","10-Q"],
"dateFrom":"2024-01-01",
"chunking":"paragraph",
"maxItems":12
}

2. Monitor risk-factor language across an industry

Full-text search for a phrase in recent annual reports — who is talking about it, and since when?

{
"mode":"search",
"query":"\"supply chain disruption\"",
"formTypes":["10-K"],
"dateFrom":"2025-01-01",
"maxItems":200
}

3. Time-series financials without parsing a single PDF

Quarterly revenue, net income and diluted EPS as clean rows, ready for a spreadsheet or BI tool:

{
"mode":"facts",
"ticker":"MSFT",
"factTags":["RevenueFromContractWithCustomerExcludingAssessedTax","NetIncomeLoss","EarningsPerShareDiluted"],
"dateFrom":"2023-01-01",
"maxItems":500
}

4. Track insider activity and ownership filings

{
"mode":"filings",
"ticker":"TSLA",
"formTypes":["Form 4","SC 13D","SC 13G"],
"dateFrom":"2026-01-01",
"maxItems":100
}

Pricing

Pay-per-event — you only pay for what gets saved:

Example: a RAG corpus of 50 full 10-K documents costs $0.25. A 1,000-row financial time series costs $0.20.

FAQ

Do I need an SEC API key?

No. SEC EDGAR is a free public data source. The Actor identifies itself with a compliant User-Agent and stays under the SEC's rate limit automatically.

How far back does full-text search go?

EDGAR full-text search covers filings from 2001-05-04 onward. Filing metadata and XBRL facts go back much further (metadata to 1994).

Why are some filings skipped in fulltext mode?

Forms whose primary document is not HTML (e.g. XBRL-only Form 4 XML) can't be section-parsed and are skipped — you are not charged for them.

Why did facts mode return the same period twice (with dedupeFacts: false)?

XBRL filings restate prior-period comparatives — a Q2 10-Q also reports last year's Q2 column. With deduplication off you get one row per filing that reported the fact (accessionNo/filed tell you which one); with the default on, only the original disclosure is kept.

Which chunking should I use for embeddings?

paragraph produces ~2000-character chunks aligned to paragraph boundaries with section attribution — ideal for most embedding models. Use section if you want one chunk per Item section.

Legal & Compliance

This Actor accesses only public data published by the U.S. Securities and Exchange Commission, which mandates open access to EDGAR. Requests are rate-limited below SEC's published threshold (10 req/s) and carry an identifying User-Agent, per SEC developer guidelines. No personal data is collected.