Insider Activity

Insider Trading Data API: Query Form 4 Filings at Scale

What a good insider trading data API looks like: Form 4 mechanics, amendment handling, derivative separation, provenance, and code examples for agents.

Updated July 11, 20266 min read
Run the Form 4 query

Pull insider trades with transaction codes intact

Use the article to understand the filing rules, then query the live Form 4 feed. The API separates open-market trades from awards and exercises, with source accessions attached.

First call
curl -H "Authorization: Bearer $ARKOLITH_KEY" \
  "https://arkolith.com/api/v1/insider/transactions?ticker=TSM&signal=true"

Agent prompt

Use Arkolith to find recent open-market insider buys for TSM, explain each transaction code, and cite the source filing.

The short version

An insider trading data API serves SEC Form 4 filings (officer, director, and 10%-owner transactions) as clean, queryable rows instead of raw EDGAR documents. The filings themselves are free at the source, so what you are actually buying is the processing: amendment supersession so corrected filings replace originals, derivative legs separated from open-market stock trades, identifiers resolved to tickers, and provenance back to the source document. Get those wrong and the "signal" you compute is noise. Arkolith serves 1.4 million+ Form 4 transactions across 7,700+ issuers over REST and MCP with those guarantees built in; broad market-data providers cover insider data too, with different trade-offs covered below.

What Form 4 data is, and why the two-day clock matters

Section 16 of the Exchange Act requires corporate insiders (officers, directors, and beneficial owners of more than 10% of a class of stock) to report transactions in their company's securities under the SEC's Section 16 guidance within two business days. That deadline is the whole reason this dataset matters: while a 13F snapshot of institutional holdings arrives up to 45 days after quarter end, a Form 4 lands while the trade is still fresh. It is the fastest legally mandated ownership disclosure that exists.

Each filing reports the insider, their relationship to the issuer, the transaction date, a one-letter transaction code (open-market purchase P, sale S, option exercise M, award A, and so on), shares, price, and post-transaction holdings, split across a non-derivative table (stock) and a derivative table (options, RSUs, warrants). The SEC's Form 4 documentation defines the full schema.

If you only need to look up one filing, the SEC's human-readable EDGAR search is free and fine. The API question starts when you want all of them, continuously, as structured rows you can filter, join, and feed to an agent.

What separates a serious insider API from a filing mirror

Anyone can mirror EDGAR XML. The engineering that makes the data trustworthy is less visible, and it is exactly where naive pipelines produce wrong answers:

Processing step Why it matters If it is missing
Amendment supersession (4/A) Insiders file corrections; the original must stop counting Corrected trades double-count; totals inflate
Derivative vs non-derivative separation An option exercise is not an open-market purchase Routine RSU vesting reads as conviction buying
Signal filtering Codes like A (award) and G (gift) are not trades Compensation events pollute buy/sell signals
Ticker and identity resolution Filers are identified by CIK, not ticker You cannot join insider activity to price or 13F data
Date plausibility guards Filer typos happen (future-dated and decade-off dates exist in real filings) MIN/MAX and window queries silently corrupt
Provenance per row An agent should cite the source accession Numbers are unverifiable; hallucination risk shifts to you

Arkolith's pipeline handles each of these explicitly. As of July 2, 2026, the serving set holds 1.42 million+ non-superseded Form 4 transactions across 7,706 issuers, with 21,000+ superseded rows retained but excluded from every aggregate (amendments replace, never double-count). Coverage is dense from today back through 2022, and a historical backfill is deepening the archive toward 2006 daily. The most recent transaction in the dataset is from today: the ingest tracks EDGAR's two-business-day rhythm.

Querying it: REST and MCP examples

One key covers both surfaces. REST for scripts and backends:

# Recent insider transactions for a ticker
curl -H "Authorization: Bearer $ARKOLITH_KEY" \
  "https://arkolith.com/api/v1/insider/transactions?ticker=NVDA"

# Only open-market buys and sells (codes P/S, the signal subset)
curl -H "Authorization: Bearer $ARKOLITH_KEY" \
  "https://arkolith.com/api/v1/insider/transactions?ticker=NVDA&signal=true"

For AI agents, the same data is a native MCP server: connect https://arkolith.com/api/mcp once and the agent calls tools like insider.transactions, insider.clusters, and signal.confluence (which crowds insider buying against 13F and Congressional activity in one call). Every response row carries provenance to its source filing, so the agent can cite the accession number instead of asserting a number from memory.

Where a broad provider fits better

Honest scoping, because no single API is best at everything. If your workflow mainly needs real-time prices, fundamentals, or broad multi-asset coverage with insider data as a side dish, a broad provider like Financial Datasets or FMP bundles all of it under one subscription, and that convenience is worth real money. The full breakdown is in our comparison of financial data APIs for AI agents.

Arkolith's lane is depth on the ownership cluster: Form 4 joined to 13F institutional holdings and Congressional trades, amendment-correct, point-in-time, with per-row provenance. If the question is "who is buying, and is the buying corroborated across filers," that join is the product. If the question is "what is the price right now," it is not.

Reading insider data without fooling yourself

Three practical rules from operating this dataset. First, filter to open-market codes (P and S) before computing any sentiment; awards, gifts, and exercises dominate raw volume and mean nothing about conviction. Our guide to tracking insider transactions walks through the codes. Second, respect the derivative split: an officer exercising options and selling the shares the same day is compensation mechanics, not a bearish signal. The derivative vs non-derivative breakdown covers the traps. Third, cluster beats single: one director buying is weak evidence, three insiders buying the same week is a different object. That is why the API exposes cluster detection as a first-class query rather than leaving every consumer to reimplement it.

If you are new to the form family, Form 3 vs Form 4 vs Form 5 explains which filing carries which event.

Frequently asked questions about insider trading data APIs

Is insider trading data legal to use?

Yes. Form 4 filings are mandatory public disclosures under Section 16, published by the SEC. Trading on your own analysis of public filings is legal; the "insider trading" that is illegal is trading on material non-public information, which this data is definitionally not.

How fresh is Form 4 data?

Insiders must file within two business days of the transaction, and filings appear on EDGAR the same day they are accepted. A pipeline polling EDGAR continuously serves trades one to three days after they happen. Arkolith's most recent transaction row is from the current day at the time of writing.

What is a Form 4/A and why does it matter for an API?

A 4/A amends a previously filed Form 4, usually to correct shares, price, or a date. A correct API must supersede the original row when the amendment arrives, or every aggregate built on the feed double-counts the corrected trade. Ask any vendor how they handle amendments before you buy; it is the single most common silent defect in insider datasets.

Can I get insider data for free?

Yes, from EDGAR directly, if you build the parser, the amendment logic, the identity resolution, and the monitoring yourself. The filings are public domain. The API price is for the processing and its maintenance, not the data. For evaluation, Arkolith's free tier includes enough credits to test the full insider toolset without a card.


Data in this article was queried from Arkolith production on July 2, 2026. Source: SEC EDGAR Form 4 filings, ingested and amendment-resolved by Arkolith. Not investment advice.

Query it yourself: get a free API key, browse insider activity by ticker, or read the API docs.

#Form 4#insider trading#API#SEC filings#AI agents#MCP#data quality