SEC Demand Signal Scraper

Apify actor that scrapes SEC EDGAR for wealth management demand signals. Identifies newly-liquid high-net-worth individuals from public filings — insider stock sales, M&A events, and intent-to-sell notices.

ICP match: First-generation wealth creators ($50M–$3B) triggered by liquidity events — the exact profile wealth management firms like TwinFocus target.

The Real-World Filter

Not all insider sales are demand signals. Meta's COO selling 2% of holdings weekly is a routine 10b5-1 plan — he already has Goldman. The CTO of an $800M company selling 60% of holdings for the first time? That's a wealth event.

Sell ratio is the #1 filter: sharesSold / (sharesSold + sharesRemaining)

Sell Ratio	What It Means	WM Signal?
>50%	Cashing out — life event	Yes (critical)
20-50%	Significant liquidation	Yes (high)
5-20%	Planned selling	Maybe
<5%	Routine trim	No (noise)

Verified against live data:

• Meta COO Olivan: sold $991K but only 2% of holdings → MEDIUM (routine)
• Meta CFO Li: sold $36.5M and 100% of holdings → CRITICAL (she liquidated everything)

Signal Sources

Filing	Signal	Timing	Example
Form 4	Insider sold stock	2 days after trade	CFO sold 60% of holdings for $5M
Form 144	Intent to sell restricted stock	Before the sale	Director filed notice to sell 177K shares
8-K (1.01/2.01)	M&A deal signed/completed	Day of event	Rallybio Corp signed definitive agreement

How It Works

1. Form 4 — Polls EDGAR live feed for recent insider filings. Fetches raw XML, parses transactions, computes sell ratio, filters out routine selling.
2. Form 144 — Same approach. Identifies insiders who filed notice of intent to sell restricted stock (leading indicator — filed BEFORE the sale).
3. 8-K — Uses EDGAR full-text search (EFTS) for M&A keywords. Classifies by item number (1.01 = deal signed, 2.01 = deal closed).
4. Scoring — Each signal scored by sell ratio, transaction value, insider seniority, and recency → urgency level.
5. Enrichment — Optional Apollo company size + domain (free endpoint). Mega-cap filter removes 10000+ employee companies.

No Playwright needed — pure API calls. No auth. Rate limit: 10 req/sec.

Input

{
  "scrapeForm4": true,
  "scrapeForm144": true,
  "scrape8K": true,
  "startDate": "2026-03-01",
  "endDate": "2026-03-09",
  "minTransactionValue": 100000,
  "minSellRatio": 0.05,
  "minUrgency": "medium",
  "officersOnly": false,
  "excludeMegaCap": true,
  "maOnly": false,
  "searchTerms": "\"acquisition\" OR \"merger\" OR \"definitive agreement\"",
  "maxForm4": 200,
  "maxForm144": 100,
  "max8K": 200,
  "maxResults": 500,
  "apolloApiKey": ""
}

Output Record

{
  "company_name": "Acme Corp",
  "company_ticker": "ACME",
  "person_name": "Jane Smith",
  "person_title": "Chief Technology Officer",
  "signal_type": "insider_sale",
  "signal_date": "2026-03-05",
  "signal_description": "Jane Smith (Chief Technology Officer) sold 45,000 shares of ACME for $5.2M — 62% of holdings",
  "signal_urgency": "critical",
  "transaction_value": 5200000,
  "transaction_value_display": "$5.2M",
  "shares_sold": 45000,
  "shares_remaining": 27500,
  "sell_ratio": 0.62,
  "sell_ratio_display": "62% of holdings"
}

Local Development

npm install
npm run start:dev    # Runs with tsx
npm run build        # TypeScript → dist/

Files

src/
  main.ts        — Actor entry point, orchestrates 3 phases + enrichment + mega-cap filter
  edgar.ts       — SEC EDGAR API client (EFTS, submissions, archives, live feed)
  parsers.ts     — XML parsers for Form 4, Form 144, 8-K classification
  scoring.ts     — Sell ratio + urgency scoring, output record builder
  enrichment.ts  — Apollo company enrichment (optional, free endpoint)

SEC EDGAR API Endpoints Used

Endpoint	URL Pattern	Purpose
EFTS Search	efts.sec.gov/LATEST/search-index?q=...&forms=8-K	Full-text search for 8-K M&A filings
Submissions	data.sec.gov/submissions/CIK{padded}.json	Company/person filing index
Archives	sec.gov/Archives/edgar/data/{cik}/{acc}/{doc}.xml	Raw filing XML
Live Feed	sec.gov/cgi-bin/browse-edgar?...&output=atom	Most recent filings (firehose)

All free. Only requirement: User-Agent header with company name + email.