SEC Insider Intelligence

One API call turns any US public company ticker into scored insider trading signals, parsed SEC filings, and a composite health rating.

Who this is for

• VC analysts screening portfolio companies for red flags before board meetings
• PE due diligence teams validating targets before term sheets
• Hedge fund researchers tracking insider behavior across positions
• Competitive intelligence firms monitoring company health at scale
• M&A advisors flagging risk patterns during deal evaluation

What you get

Enter a ticker or company name. The actor returns three row types in a single dataset:

Row type	Contents
filing	SEC filing metadata with 8-K event classification
insider_trade	Parsed Form 4 data: insider name, title, shares, price, total value
signal	Scored intelligence signal (0-100) with direction, confidence, and narrative

A health score summary is stored in the key-value store under the SUMMARY key.

The 5 intelligence signals

1. Filing Velocity

Quarter-over-quarter filing trend. Acceleration often precedes major corporate events -- new products, restructurings, acquisitions. Slowdown may signal operational retreat. Direction: bullish or bearish.

2. 8-K Event Risk

Detects high-risk 8-K items: leadership departures, financial restatements, material impairments, delisting notices, auditor changes. More events and higher severity produce higher scores. This is the earliest public signal of corporate distress.

3. Insider Net Sentiment

Dollar-weighted buy/sell ratio from Form 4 filings. Insider buying is the strongest bullish indicator in public markets -- executives put their own money at risk. Heavy net selling tells the opposite story.

4. Insider Concentration

Flags when a single insider accounts for 40%+ of total sell volume. C-suite concentrated selling scores highest. When the CFO sells more than the rest of the board combined, pay attention.

5. Insider Selling Before Events

Cross-references insider sales against high-risk 8-K filings within a 90-day window. Three insiders sold $2.8M of GE stock before 8-K risk events. This sell-before-event pattern is what regulators watch most closely.

Sample output

Signal row

{
  "row_type": "signal",
  "source": "intelligence",
  "company_name": "GENERAL ELECTRIC CO",
  "signal_name": "Insider Selling Before Events",
  "signal_category": "cross_reference",
  "signal_score": 90,
  "signal_direction": "alert",
  "confidence": "high",
  "data_points": 3,
  "narrative": "3 insider sale(s) totaling $2,789,435 occurred within 90 days before high-risk 8-K filings. This sell-before-event pattern is closely watched by regulators.",
  "supporting_evidence": [
    "Stokes Russell sold $2,381,680 on 2025-11-19, 57d before 8-K on 2026-01-15",
    "Ali Mohamed sold $306,701 on 2025-08-05, 57d before 8-K on 2025-10-01"
  ]
}

Insider trade row

{
  "row_type": "insider_trade",
  "source": "sec_edgar",
  "company_name": "GENERAL ELECTRIC CO",
  "cik": "0000040545",
  "filing_date": "2026-02-03",
  "insider_name": "Stokes Russell",
  "insider_title": "Senior Vice President",
  "is_officer": true,
  "transaction_type": "sale",
  "transaction_code": "S",
  "shares": 8894.0,
  "price_per_share": 306.51,
  "total_value": 2726099.94,
  "shares_owned_after": 150434.0,
  "transaction_date": "2026-01-30"
}

Tested on 20 companies

Results from real runs against SEC EDGAR (2-year lookback, default settings):

Company	Filings	Insider Trades	Signals	Health Score
Apple	30	38	4	31/100 (warning)
Tesla	30	72	5	62/100 (stable)
Boeing	30	33	4	80/100 (strong)
NVIDIA	30	181	4	17/100 (critical)
Snowflake	30	92	5	25/100 (warning)
GE	30	29	5	41/100 (caution)

NVIDIA's 181 insider trades and 17/100 health score tell a clear story. Boeing's 80/100 shows the signals can read both directions. The data speaks for itself.

Input reference

Field	Required	Default	Description
companyName	Yes	--	Ticker or company name ("AAPL" or "Apple Inc")
includeInsiderTrades	No	true	Parse Form 4 insider trades. Set false for faster, cheaper filing-only runs.
dateFrom	No	2 years ago	Start date (YYYY-MM-DD)
dateTo	No	Today	End date (YYYY-MM-DD)
maxFilingResults	No	200	Maximum filing rows (max 1000)
formTypes	No	All major types	SEC form types to include
secUserAgent	No	Default agent	SEC EDGAR requires a User-Agent with name and email
proxyConfig	No	None	Apify proxy settings

Default form types: 8-K, 10-K, 10-Q, Form 4, SC 13D/G, S-1, DEF 14A (and amendments).

Pricing

$0.005 per result row (pay-per-event).

Scenario	What you get	Rows	Cost
Quick scan (filings only)	Filing metadata + 8-K classification	~35	~$0.18
Standard run	Filings + insider trades + all 5 signals	~70	~$0.35
Large-cap deep dive	Full analysis, high-activity company	~200	~$1.00

Set includeInsiderTrades to false for filing-only runs. This skips Form 4 parsing and disables the three insider signals, cutting cost and runtime.

How the health score works

The composite score (0-100) is a weighted average of all generated signals. Each signal's weight equals its data point count -- signals backed by more evidence count more.

• Bullish signals contribute their score directly (higher = healthier)
• Bearish signals contribute the inverse (high bearish score = low health)
• Alert signals penalize the score (sell-before-event patterns, concentrated selling)
• Neutral signals anchor toward 50

Score labels: critical (0-19) | warning (20-39) | caution (40-59) | stable (60-79) | strong (80-100)

The health score is stored in the key-value store under SUMMARY alongside an executive summary and top signal narratives.

Data source

All data comes from SEC EDGAR -- the official public filing database of the US Securities and Exchange Commission. No paid data feeds. No scraped paywalled sources. Free, official, public records.

The actor resolves CIK numbers automatically from either a name or ticker. Form 4 XML documents are parsed directly for insider trade details that the EDGAR JSON API does not expose.

Limitations

• US public companies only. Covers SEC-registered companies. No foreign-only or private companies.
• Non-derivative Form 4 transactions only. Derivative transactions (options, warrants) are not yet parsed.
• SEC rate limit. EDGAR allows 10 requests/second. Large runs take 30-60 seconds.
• Pattern detection, not prediction. Signals are quantitative patterns from filed data. They are not buy/sell recommendations.
• Historical coverage depends on filings. If a company hasn't filed it, the actor can't find it.
• Not investment advice. This is a research tool. Use it as a starting point for your own analysis.