The problem
Every trading day, between 200 and 600 SEC Form 4 insider-transaction filings hit EDGAR. The vast majority are noise: routine 10b5-1 sales, mechanical option exercises, automatic award grants. The few filings that do carry information — a CFO's first open-market buy in three years, three officers piling into the same ticker the same week, a director quietly doubling their position into a beaten-down stock — get buried in the rest.
Free tools like OpenInsider dump the raw data. Bloomberg charges $24,000 a year to interpret it. Retail traders sit in the middle, knowing the signal exists but unable to extract it from the firehose.
That gap is what Form 4 AI is built to close.
What Form 4 AI does
Every filing that hits EDGAR — typically within 60 seconds of publication — is parsed, scored, and written up as a structured analyst thesis. Each thesis is grounded in four hard inputs:
- Trade size in context of the insider's prior compensation and personal holdings.
- Insider role — CEO, CFO, board member, 10% holder. We treat sole-officer buys as much higher-information than 10%-holder trims.
- Historical edge — every insider's 90-day, 1-year, and 3-year forward edge vs SPY computed from their own prior open-market buys.
- Surrounding context — recent 8-K filings, price action, cluster detection (≥3 insiders in 7 days = high-signal pattern).
Out the other side: a calibrated paragraph, a 0–100 conviction score, a key-factor breakdown, and a list of risks. Decisions in 30 seconds instead of 30 minutes per filing.
What you get for free
The free tier is a real product, not a teaser. You see:
- The last 30 days of filings, all scored.
- Conviction-score badges on every filing.
- Public insider and issuer pages.
- A weekly Top-5 digest by email.
- The full public track record page with forward returns vs SPY for every conviction-≥70 call.
Pro at $15/month unlocks the full archive, the AI thesis paragraph on every filing (not just the score), watchlists, and email/SMS alerts on conviction thresholds. Premium at $49/month adds Opus-tier deep analysis on flagged buys and SMS dispatch within seconds of EDGAR publication.
Why a public track record
"Conviction 87" is meaningless unless the model is willing to be graded in public. So we grade ourselves: every conviction-≥70 thesis is auto-tracked through 7-, 30-, and 90-day forward returns benchmarked against SPY. Updated nightly. No cherry-picking — the calls that didn't work are on the same page as the ones that did.
With fewer than 30 settled calls in any direction, treat aggregate numbers as directional. Insider-buy backtests are notoriously noisy and notoriously path-dependent. We publish raw counts, not just averages, so you can apply your own significance bar.
How this is different from the alternatives
OpenInsider and Quiver dump raw data and basic filters. We layer interpretation on top. WhaleWisdom focuses on 13F holdings, not Form 4. Bloomberg does insider analysis well — at $24k/year. We sit at the intersection most people actually want: institutional-quality interpretation at retail-affordable pricing, with a public scoreboard.
What's next
Over the coming weeks, we're shipping:
- A public daily Substack with the day's top conviction-≥70 call.
- An API tier for traders who want to plug Form 4 AI into their own workflow.
- Backfilled forward-return numbers on the track-record page.
- Cluster-buy SMS alerts (Premium) within seconds of the third filing landing.
If you want to follow along, the email signup at the bottom of this page is the best way. One email a week, real value, no spam. Reply with feedback — every email is read by a human.
Disclaimer
This is journalism over public regulatory filings, not investment advice. AI-generated theses are pattern recognition on disclosed data, not stock picks. Past insider performance doesn't guarantee future returns. Insider-buy backtests are notoriously noisy and path-dependent. Do your own work.