How can we help you?

Insider Intel is a financial intelligence platform that tracks SEC Form 4 insider trading filings in real time. When a company director, executive, or major shareholder buys or sells shares in their own company, they are legally required to report it to the SEC within two business days. Insider Intel collects those filings the moment they appear, scores them using a six-factor AI model, and presents the results in an actionable way for individual investors.

The platform is built for retail and self-directed investors who want to know what company insiders are doing with their own money — without having to read raw SEC filings.

Go to the Charts view and type a ticker symbol in the search field. When the chart loads, click the heart (follow) icon near the company name to add it to your Company Watchlist. You can also follow companies directly from the Dashboard Transactions feed or the Buy Activity view.

• Dashboard — market-wide view. Recent insider transactions, live buy/sell ratio, and your watchlist with AI Signal scores.
• Charts — single-company view. Price chart overlaid with insider trade markers, trades table, and full AI analysis.
• AI Insights — personalised daily summary. AI-generated write-ups for each company on your watchlist.

Go to Monitor → Alert Rules and click Create rule. Choose trigger conditions: a new insider buy or sell for a specific ticker, a threshold number of purchases, a change in AI Signal level, or a combination. Manage delivery preferences in Settings → Manage Preferences.

A full guided tour is on our roadmap for the beta launch. In the meantime, the Why Us page explains the six-factor AI Signal model, and the Roadmap page shows where the product is heading.

AI Outlook is a single-input score from Claude AI's qualitative analysis of recent insider filings — it reflects the AI's read of insider behaviour on a scale of 0–100 (50 = neutral). It is based on the most recent filings on record for that company.

AI Signal is a composite score from six weighted factors: insider buying activity (25%), analyst consensus (25%), price upside vs. target (20%), AI Outlook (up to 20%), insider track record (15%), and breakout pressure (10%). It drives the Accumulate / Watch / Hold / Caution / Avoid labels and is the more actionable of the two. Learn how it's built →

When the scores diverge, that gap is itself informative — it often reflects a tension between the qualitative AI read and what the quantitative signals are showing.

• Accumulate — strong positive signal across multiple factors (score ≥ 84/100).
• Watch — positive but not yet compelling; one or more factors strong (68–83).
• Hold — neutral; no strong signal in either direction (52–67).
• Caution — some negative factors present (36–51).
• Avoid — multiple negative factors aligned (below 36).

These labels are informational, not financial advice.

• Insider buying activity — 25%. Recent open-market purchases weighted by size, role seniority, and number of distinct insiders. A cluster bonus multiplier is applied when 3 or more distinct insiders buy within a 4-day window: 3–4 insiders ×1.15, 5–6 insiders ×1.25, 7 or more insiders ×1.35. This elevates coordinated buying — where multiple executives act in the same short window — above staggered individual purchases of equivalent total value.
• Analyst consensus rating — 25%. Aggregate analyst recommendation from Finnhub, scaled by a coverage depth discount: the raw rating score is multiplied by min(analyst count ÷ 5, 1.0), so a Strong Buy backed by 1 analyst carries proportionally less weight than the same rating backed by 5 or more analysts. Full weight is reached at 5 analysts.
• Price upside vs. analyst target — 20%. Gap between current price and mean analyst price target.
• AI Outlook — up to 20%. Weight scales with confidence: High = 20%, Medium = 15%, Low = 0%.
• Insider track record — 15%. Historical performance of insider purchases for this company.
• Breakout pressure — 10%. Whether the chart shows a live ascending triangle: price coiling under a flat ceiling it has tested several times, with each pullback bottoming higher, or just breaking through. It only counts when a setup is live: if there is no pattern, or it has gone stale, the factor is left out and the other factors carry its weight. A failed pattern scores low.

When multiple insiders buy in a tight window, that coordination is a qualitatively stronger signal than the same total dollar value spread across isolated purchases over 30 days. A CEO buying $1M on Monday and a CFO buying $800K on Wednesday suggests a shared internal conviction that a standard per-trade count would not distinguish from two unrelated buys a month apart.

The cluster bonus is a multiplier applied to the Insider Buy Activity score (25% of the AI Signal composite) whenever 3 or more distinct insiders make open-market purchases within any 4-day rolling window in the 30-day lookback period:

• 3–4 insiders in 4 days — ×1.15
• 5–6 insiders in 4 days — ×1.25
• 7 or more insiders in 4 days — ×1.35

The multiplier is applied before the factor's 5.0 ceiling, so a strong pre-bonus score can still be elevated meaningfully. The cluster details — insider count, window span, and the multiplier applied — are shown in the score breakdown drawer on the Charts page.

Only individual insiders count toward the cluster threshold. Institutional investors (funds, partnerships, and LLCs) are excluded from the AI Signal scoring pipeline entirely, so a coordinated cluster reflects genuine executive conviction rather than a single large fund purchase split across filings.

A consensus rating is only as reliable as the number of analysts behind it. A single analyst calling a stock a Strong Buy is meaningful, but it carries far less evidential weight than ten analysts independently reaching the same conclusion. The coverage depth discount reflects this: the raw rating score is multiplied by min(analyst count ÷ 5, 1.0).

• 1 analyst → ×0.20 (20% of full weight)
• 2 analysts → ×0.40
• 3 analysts → ×0.60
• 4 analysts → ×0.80
• 5 or more → ×1.00 (full weight)

The discount is visible in the score breakdown drawer on the Charts page — the Analyst Consensus card shows the base rating, the depth discount applied, and the resulting adjusted score. If no analyst count data is available, no discount is applied.

No. Institutional investors — funds, partnerships, LLCs, and other entities identified as non-individual filers — are excluded from the AI Signal scoring pipeline entirely. Only individual insiders (executives, directors, and named officers) contribute to the Insider Buy Activity score and the cluster buy detection.

This matches the behaviour in the Charts and Dashboard views, which also filter out institutional trades and display a notice showing how many were excluded. The AI Signal table reflects the same population — individual conviction buying only.

The reason institutional trades are excluded is that a fund acquiring a large position through Form 4 filings does not carry the same signal as a CEO or CFO making a discretionary open-market purchase with personal capital. Including institutional volume would distort the buy value normalisation that ranks all tickers relative to each other.

Confidence measures data richness, not model accuracy. High — multiple recent filings with strong insider activity; AI Outlook carries its full 20% weight. Medium — reasonable data with some gaps; weight reduced to 15%. Low — limited data; AI Outlook shown for reference but carries 0% weight in the composite.

Freshness cutoff. An AI Outlook score older than 30 days is treated as unavailable regardless of its confidence level — it is excluded from the composite entirely and the remaining five factors renormalise to fill the gap. The score is still displayed in the drawer for reference, labelled as stale, but it does not influence the AI Signal. This prevents a pre-sector-shock analysis from quietly pulling the composite in the wrong direction after conditions have changed.

They measure different things. AI Outlook is one qualitative input; AI Signal combines five other quantitative factors alongside it. A positive AI Outlook (strong insider buying pattern) can coexist with a lower AI Signal if analyst ratings are weak or price upside is limited. The divergence is information, not an error — it shows where the qualitative and quantitative reads disagree.

AI Signal is recalculated whenever a new Form 4 filing arrives for a tracked company, and on a daily cycle for all watchlisted companies. AI Outlook is refreshed daily — the timestamp in the Charts view shows when the current version was generated.

Three things update automatically at different cadences once an insider trade is on record:

1. Breakout Pressure score — refreshed daily. This factor reads the stock's daily price chart for a live ascending triangle: a flat ceiling the price has tested several times, with each pullback bottoming higher. It scores highest as the price squeezes up against the ceiling, or has just broken through it.

   So a stock coiling 2% below its ceiling scores higher than the same stock 10% below it. If there is no live pattern the factor simply isn't counted, and it scores low if the price breaks below its rising support. Stocks on a watchlist or with insider buying are refreshed every day; others every few days.

2. Insider track record score — updated every 4 hours. Every executive with open-market purchase history carries an ongoing performance grade (A / B / C / D / E) based on how those prior trades are performing at current prices — essentially their win rate (% of buys currently in profit) and average unrealised P&L.

   If a CEO's prior purchases are all in profit after a stock gain, their Grade and win rate rise and that track record score (15% of the AI Signal composite) updates accordingly. A CEO with a Grade A and 80% win rate carries meaningfully more weight than one with no track record.

3. AI Outlook narrative — when the AI regenerates a company's summary, it receives each filing executive's current track record alongside the raw filings.

   It can then reason explicitly: "this CEO's prior open-market purchases are broadly profitable at current prices, consistent with a pattern of well-timed conviction buying" — as opposed to treating each new purchase in isolation.

Future: the system tracks unrealised P&L today (current price vs. trade price) but not realised exits. A closed-trade batting average — knowing a trade was closed at +40% after 90 days — is a planned future enhancement built on the Signal Outcome data the platform is already accumulating.

The current AI Signal model uses a fixed six-factor weighted composite. It is deliberate and interpretable by design — every weight can be explained and reasoned about. As the platform accumulates more historical data, we plan to evolve the model in three stages:

• Factor Correlation Analysis. Using real historical outcomes (how a stock performed 30, 60, and 90 days after an insider buy) to validate whether each of the six factors is actually predictive, and to identify which combinations of factors have the highest correlation with positive returns. This will inform weight adjustments grounded in evidence rather than intuition.
• Ensemble scoring. Layering a secondary model alongside the existing composite — for instance, a gradient-boosted classifier trained on the historical outcome data — and surfacing both scores so users can see where the two approaches agree or diverge.
• Dedicated neural network. Once sufficient labelled outcome data has been accumulated (target: 12–18 months of live production data), we plan to evaluate training a purpose-built neural network on the full feature set: filing metadata, price action, analyst signals, sentiment, and macro context. The goal is a model that can identify non-linear interactions between factors that a fixed-weight composite cannot capture.

All of this work is tracked on the public roadmap. We are committed to keeping the model explainable even as it grows more sophisticated — users will always be able to understand why a stock has the score it has.

SEC Form 4 filings must be submitted within two business days of the trade. Insider Intel polls the SEC EDGAR feed continuously during market hours and processes new filings within minutes — typically 15–30 minutes after acceptance by EDGAR. The "Filings live" indicator pulses green when the real-time feed is active.

• Insider filings — SEC EDGAR (the official US SEC filing system).
• Stock prices & history — End-of-day OHLCV price data sourced from a third-party market data provider.
• Analyst ratings & price targets — Finnhub.
• AI Outlook analysis — Claude AI (Anthropic), processing recent insider filings and trading patterns.

Prices reflect the previous day's closing price, updated each morning. They are not live streaming quotes. For live intraday prices, cross-reference your broker or a live market data service.

Filings and prices update on different schedules. A Form 4 filed at 18:30 ET appears within minutes, but the closing price for that day may not be available until the following morning. The filing and chart align once the next price update runs.

The countdown shows how long until the next scheduled full data pull from SEC EDGAR. The green pulsing "Filings live" dot indicates whether the real-time feed is active — new filings typically appear well before the countdown reaches zero.

The trade filter sits above the trades table. Buy, Sell, and All Trades are independent toggles. On initial load the filter auto-selects the most relevant view; if a stock only has insider sells in the selected period, it switches to "All Trades" so the table isn't empty. Manual adjustments are respected for the rest of the session.

• The selected period has no filings. Try 1Y or 2Y — insider activity clusters around earnings windows and can be absent for months.
• The filter has no matches. If "Buy" is selected but the company only has insider sells, switch to "All Trades".
• The stock recently listed. Filings only exist from the listing date onward.

A "special trade" is where the filing price differs significantly from the market price — typically an options exercise, gift transfer, or Rule 10b5-1 plan trade. These are shown in the table for completeness but excluded from AI Signal weighting. A CEO exercising options at $0.01 is not the same signal as spending $500,000 on the open market.

The same exclusion applies in the Insider Screener: any trade where the filed price deviates more than 25% from the market closing price on the trade date is excluded from that insider's Return rate and Win rate calculations. This prevents distorted pricing — such as ADR filings that report a foreign-exchange share price rather than the US market price — from inflating or deflating an insider's score. The trade still counts toward their Conviction and Recency factors, as the purchase itself remains on record.

• Green upward arrow — open-market purchase (insider buy).
• Red downward arrow — open-market sale (insider sell).
• Grey marker — special trade (options exercise, gift, or other non-open-market transaction).

Click any marker to see full trade detail. The corresponding trades table row is highlighted when you hover a marker.

Calculated as (current price − trade price) × shares for each open-market purchase still held. Uses the previous close as the current price. Does not account for dividends, stock splits, or tax — an indicative figure for context, not a precise P&L.

The table shows filings whose filed date falls within the selected window. Insider trading is inherently clustered — many companies go months between any activity. An empty 3M window does not mean data is missing.

A transaction where the insider bought shares on the stock exchange at the prevailing market price — the same way any retail investor would, identified by transaction code P on the Form 4. It is the strongest insider signal because the insider is voluntarily spending their own money at market rates with no guaranteed return.

Under Section 16 of the Securities Exchange Act: any officer, director, or beneficial owner of more than 10% of a class of registered equity securities. They must file Form 4 within two business days of any transaction. The strongest signals typically come from C-suite executives, board directors, and large shareholders.

The period during which insiders are permitted to trade under the company's internal trading policy. Windows typically open shortly after quarterly earnings and close before the next earnings period. They are shown as shaded regions on the Charts price chart.

The SEC filing that reports changes in beneficial ownership by insiders. It discloses: the insider's name and role, transaction date, type, number of shares, price per share, and total shares held after the transaction. Each row in the Insider Intel trades table corresponds to one transaction line from a Form 4.

Insider Intel classifies roles into 16 standard categories by keyword-matching the SEC filing's officerTitle field. Filers use inconsistent formats so the displayed role may be an approximation. Hover over the role label in the trades table to see the exact title from the original Form 4.

The Insider Screener ranks every individual insider who has filed an open-market purchase (Form 4 code P) by their historical track record. It answers the question: which insiders consistently buy at the right time?

Each insider is scored 0–100 and assigned a grade A–E. The score is designed to reward both insiders whose positions are currently profitable and insiders who timed their entry well — even if the stock has since given back some of its gains. Scores update every 4 hours.

Only open-market purchases are used — grants, awards, and Rule 10b5-1 plan exercises are excluded, as those are not genuine expressions of conviction. Institutional entities (funds, limited partnerships, LLCs, and similar) are also excluded — see Why are institutional investors excluded? below.

The score is a weighted composite of five factors, each normalised to a 0–5 scale before weighting:

The 90-day peak blend in Return rate and Win rate is deliberate: an insider who bought at exactly the right moment — and whose stock subsequently ran 30% before retracing — should score well, even if the position is now flat or slightly negative. Current-price-only scoring would unfairly penalise good timing.

Price-discrepancy exclusion. Any trade where the filed price deviates more than 25% from the market closing price on the trade date is excluded from Return rate and Win rate. This prevents distorted filings — such as ADR shares reported in a foreign currency rather than the US market price — from skewing an insider's score. The trade still counts toward Conviction and Recency.

Hover over any score pill in the screener to see the full breakdown for that insider, including both their current and peak P&L figures.

Grades bucket the 0–100 score into five tiers:

Grades update automatically every 4 hours as prices move and new filings arrive. An insider's grade reflects their long-term pattern of timing, not a single trade.

Under SEC rules, any investor who holds more than 10% of a company's shares must file a Form 4 whenever they trade — including funds, limited partnerships (L.P.), LLCs, and other institutional entities. This means they appear in the same EDGAR feed as human corporate insiders.

However, institutional entities represent fundamentally different signals. A fund manager accumulating a position is making a portfolio allocation decision — not an insider at the company with first-hand knowledge of its operations, strategy, or near-term outlook. Mixing them with individual officers and directors dilutes the quality of the signal and distorts the screener rankings (a fund making 50 individual purchases across five tickers would otherwise score near the top of every metric).

For this reason, Insider Intel automatically excludes any filer whose name matches common institutional patterns — limited partnerships, LLCs, funds, management companies, advisory firms, and similar entities. This exclusion applies across:

• Insider Screener — institutional entities do not appear in the ranked list or receive a score
• Insider Watchlist — only individual insiders can be tracked
• Chart view — institutional trades are hidden from the "Insider trades in view" table; a notice at the bottom of the table tells you how many were filtered out for that stock
• Dashboard & Buy Activity — institutional purchases are excluded from the filings feed and all P&L calculations

If you specifically want to see 10% owner filings from funds or partnerships for a given stock, you can view them directly on EDGAR.

Two factors can inflate scores, particularly in rising markets:

1. Short track records benefit from survivorship. An insider who made one well-timed purchase will have a 100% peak win rate and a high return score — even though a single trade is not a statistically meaningful sample. The score cannot distinguish between genuine skill and luck until there are enough trades to form a pattern.
2. Broad market tailwinds lift all boats. In a sustained bull market, most purchases will have peaked above their cost basis within 90 days simply because markets were rising — not because the insider had edge. This inflates peak win rates across the board.

The practical fix: use the Min buys filter set to 3 or higher. An insider with five or more open-market purchases has demonstrated a repeatable pattern rather than a single lucky entry. The more trades in the sample, the more the score reflects genuine skill.

• Search — type any part of an insider's name to filter the list in real time.
• Min grade — show only insiders at or above a chosen grade tier (e.g. "B or better" hides C, D, and E).
• Min buys — minimum number of open-market purchases on record. Setting this to 3 or more is the single most effective way to filter out lucky one-trade insiders and surface statistically meaningful track records.
• Column headers — click any column header to sort ascending or descending. Score is the default sort. Click again to reverse the direction.
• Score hover — hover over any score pill to see the full breakdown: all five component scores (0–5 each), plus both current and 90-day peak P&L and win rate figures for that insider.
• Ticker pill — click the exchange:ticker pill in any row to navigate directly to that company's price chart and filing history.
• Heart icon — add an insider to your Insider Watchlist directly from the screener. You will be notified when they next file a Form 4.

An alert rule defines a condition that, when met, sends a notification. Conditions include: a new insider buy or sell for a specific ticker; a threshold number of purchases in a rolling period; a change in AI Signal level; or a combination. Delivery is configured in Settings → Manage Preferences.

Alert priority tiers (red / amber / green) are coming in the next sprint — Red = action required, Amber = watch closely, Green = for information. The level will appear in the email subject line and in-app notification. Track progress on the public roadmap.

Currently by email. Two modes in Settings → Manage Preferences: Per-alert — sent immediately when a rule fires. Batched digest — collected and delivered at your chosen times (up to three per day). Push notifications and additional channels are planned.

• Company Watchlist — companies (by ticker) you are following. Appear in AI Insights and Dashboard with AI Signal scores.
• Insider Watchlist — individual insiders you are tracking across time and companies. When a tracked insider files a Form 4 anywhere, it surfaces in the Insider Watchlist view.

It should appear within a few seconds. If not, try refreshing the page — the watchlist is loaded fresh on each page load. If it persists, contact us.

Insider name search is coming in the next release. Currently, add insiders from the Charts view (click their name in the trades table) or from the Analyse section.

Insider Intel is currently in beta and free to access for all beta users. Full pricing will be published ahead of the general launch — early adopters will be offered preferential rates. See the pricing section on our homepage for the latest details.

Insider Intel is currently in beta. Pricing details will be published ahead of the full launch — visit the pricing page for the latest information. Beta users will be notified directly and offered early-adopter terms.

Go to Settings → Manage Preferences to configure alert delivery mode, digest schedule, Daily Briefing opt-in, and notification email address.

Go to Settings → Account. If you cannot log in, use the password reset link on the login page or contact us.

Use the Contact page. During beta we aim to respond within one business day. For bug reports, include the ticker, the view you were on, and what you expected vs. what happened.