Weekly forensic screen
200+ microcap names filtered for termination risk (officer-share band), board quality (prominent backer presence), and insider cluster buying (open-market buys, multi-insider, 180-day window).
All use cases
Forensic microcapsLive in portalsign-in
Penny Picks
Forensic microcap research. 200+ tickers screened weekly for dilution, runway, and gatekeeper red flags — with cited deep-dive on any name.
Penny Picks is forensic research for the microcap universe — the part of the market where retail investors get burned most often by dilution, undisclosed conflicts, and gatekeeper-syndicate patterns. A weekly screener filters 200+ tickers across termination risk, board quality, and insider cluster buying. Click any name for a 16-node forensic deep-dive with cited evidence — and a tier-labeled verdict that's honest about what it does and doesn't know.
What makes it different: 16-node forensic framework. Verdicts grade evidence honestly — validated drivers vs. mechanism-only commentary vs. tested-null context.
What it does
Capabilities
16-node forensic dossier per ticker
LLM agent walks a structured framework: dilution history, runway months, gatekeeper syndicates, insider trades, board composition. Every finding is cited with a verbatim quote and source.
Two-sided forward outlook
For every published verdict, probability of bad outcome AND probability of >2.5× return at 1y/2y/3y. Avoid buckets show 'upside not measured' rather than fabricating an outlook.
Three-tier trust model
Validated drives the verdict (backtested). Mechanism-only is cited LLM commentary. Tested-null is structured context that didn't predict in backtest. Each labeled visibly.
How it works
Input → output
Input
WGRX · Wag! Group Co.
- 1Pull weekly screen results: termination flag, board quality, insider activity (all from SEC Form-4 + EDGAR)
- 2Render verdict pill (avoid / watch / null) with the upstream's published two-sided outlook
- 3On-demand deep-dive: LLM walks the 16-node framework, citing evidence per node
- 4Each node graded: validated (drives verdict), mechanism-only (commentary), tested-null (context)
- 5Context block surfaces SEC insider/board/network facts — but explicitly NOT a driver of the verdict
Output
Verdict pill + outlook bar + 16-card forensic dossier. Each card: question, answer (with abstention support), severity, timing, cited evidence, grade.
Output preview
portal · live
WGRX
Wag! Group Co. · microcap
Avoid
Forward outlook
by 1y44% bad
by 2y49% bad
by 3y67% bad
Upside not measured for avoid buckets
Top findings · 16 nodes walked
- Q-DIL-001yes · fatalvalidated
- Q-RUN-001yes · strongmechanism-only
- Q-OWN-001— · weakmechanism-only
Source: EDGAR · Finnhub · provenance traced
Why this isn't a wrapper
Why this isn't 'AI stock screening'
Microcap forensics is where signal honesty matters most — and where most tools fail.
- Three explicit trust tiers. We tell you which signals drive the verdict and which are just commentary — most tools blur this.
- Q-OWN-001 abstention: when no insider-ownership figure is sourced, the answer renders '—' (a known unknown), not a fabricated number.
- Coverage notes rendered verbatim per upstream's legal requirement — 'absence does NOT mean clean' framing applied throughout.
- Public-record framing on every named entity — association, never accusation.
Ready to see it?
Sign in to the portal and pull a dossier on any ticker in the universe.