Last updated: August 8, 2026

About InsiderCluster

InsiderCluster tracks one specific, public, under-watched signal: when several corporate insiders buy their own company's stock at around the same time.

Every officer, director and 10% owner of a US-listed company has to report their trades to the SEC on Form 4, usually within two business days. Those filings are free and public. The problem is volume — thousands are filed every week, and the overwhelming majority are routine compensation events, not decisions. The interesting ones are buried.

This site does the sorting. It reads Form 4 filings from SEC EDGAR, keeps only genuine open-market purchases, groups them by company, and surfaces the cases where two or more distinct insiders bought the same stock inside a 90-day window — a cluster buy. Then it tracks what happened next, measured against the S&P 500, and publishes the result whether it flatters the signal or not.

Who publishes this

InsiderCluster is published by NexarionFlux AI LLC, an independent US company. It is not affiliated with, endorsed by, or acting on behalf of the SEC, any exchange, any broker, or any company covered on the site.

We are a small independent operation, not a newsroom. That shapes what we do and don't claim: we are good at parsing filings and being explicit about our arithmetic, and we do not offer research coverage, price targets, ratings or recommendations on individual companies.

Questions, corrections and data disputes go to frank.yang@nexarionflux.ai — see the contact page for what to expect.

Why it exists

Insider selling is noisy. An executive might sell to pay a tax bill, diversify, or buy a house, and none of that says anything about the business. Buying is different: an insider already earns salary, bonus and equity from the company, so choosing to put more of their own after-tax money in is a costly, voluntary, publicly accountable bet.

A cluster of such buys is harder still to explain away. When the CFO and two directors each independently buy in the same few weeks, the simplest reading is shared conviction. That is a real, documented phenomenon in the academic literature on insider trading — and it is visible to anyone willing to read enough filings.

Most places that surface this data either put it behind an expensive terminal or bury it in an undifferentiated feed. We wanted the clusters themselves, dated, scored, and honest about their track record. That's the whole product.

What's on the site

• Cluster-buy screener — live clusters ranked by number of distinct buyers and total dollars committed.
• Company dashboards — per-ticker Form 4 history alongside price, news and social sentiment.
• Weekly leaderboards — the biggest fresh cluster formations, archived by week.
• Track records — how past clusters actually performed versus the S&P 500.
• Guides — plain-English explainers on reading Form 4, what transaction codes mean, and how to evaluate a cluster.
• Methodology — the exact definitions and arithmetic behind every number above.

Our policy on automated and AI-assisted content

We use software to do things software is good at, and we tell you when we have.

• The data is computed, not written. Filing counts, cluster thresholds, returns and benchmark comparisons come straight from the pipeline described in our methodology. No language model touches a number.
• The guides are written by people. The explainers in the blog are researched and written by hand, and edited before they go up. They are the site's editorial voice.
• The daily market briefs are machine-generated, and labelled as such. Each one summarises a single session's most-active stocks and the headlines behind them. The tables are raw data; the surrounding commentary is written by a language model working only from that data and those headlines. Every brief carries a notice saying so, and they are excluded from search indexing because they are a convenience log, not original reporting.

We do not publish machine-written articles as if a person wrote them, and we do not generate pages just to have more pages.

Accuracy and corrections

We work from primary sources — SEC EDGAR for filings, exchange data for prices — but parsing thousands of filings is not error-free, and upstream data can be wrong or revised. If something on this site looks incorrect, please tell us. We fix confirmed errors promptly, and we would rather hear about a mistake than not.

The methodology page documents the known limitations of the approach, including the ones that make the signal weaker than it might look.

How we make money

InsiderCluster is free to use. It is supported by display advertising, and we may add optional paid features later.

Advertising does not influence what appears on the site. Ads are served by third parties and are not selected, endorsed, or reviewed by us, and no advertiser has any input into which clusters we surface or how we score them. Ads are labelled and visually separated from content. See our privacy policy for what advertising means for your data, and our cookie policy for how to control it.

Not investment advice

Nothing here is investment advice, a recommendation, or an offer to buy or sell anything. Insider buying is one input among many, insiders are often early, and they are sometimes simply wrong. Do your own research and consider talking to a licensed adviser. See the full disclaimer.