Skip to content
FalsifyLab FalsifyLab Alpha MCP

insider_buy_clusters

insider_buy_clusters

SEC Form 4 cluster-buy detector. Surfaces tickers where 3 or more insiders bought within a 24-hour window. Cluster buys are higher-signal than single-insider trades.

Signature

insider_buy_clusters({
  min_insiders?: number,   // default 3
  window_hours?: number,   // default 24
  min_dollar_amount?: number,  // USD, default 100_000
  include_directors_only?: boolean,  // default false (require officer/10%-owner)
  limit?: number,          // 1-100, default 10
})

Returns

[
  {
    "ticker": "ACME",
    "company_name": "Acme Corp",
    "n_insiders": 4,
    "total_purchase_usd": 1_850_000,
    "window_start": "2026-05-12T14:30:00Z",
    "window_end": "2026-05-13T14:30:00Z",
    "buyers": [
      {"name": "Jane Doe", "title": "CEO", "shares": 5000, "usd": 425_000},
      {"name": "John Smith", "title": "CFO", "shares": 4500, "usd": 382_500},
      ...
    ],
    "edgar_urls": ["https://sec.gov/..."]
  }
]

Example agent prompt

“show me any small-cap stocks where 3+ insiders bought >$500k total in the last 24 hours.”

Why this exists

Single insider buys are noisy (10b5-1 plans, automatic stock awards). Cluster buys are intentional: officers coordinating allocation = stronger signal. The tool parses raw Form 4 XML, extracts transaction codes, filters for P (open-market purchase), groups by ticker, applies window.

Edge studied in academic lit (Cohen, Malloy, Pomorski 2012): cluster buys outperform single buys by ~6% annually.

Free tier

Today’s clusters, 10 results. Good for daily “what insiders are betting on” digest.

Pro tier

Real-time (alerts within minutes of filing). 90-day history. Filter by sector, market cap, individual title. Backtest cluster signals on historical data.

See also