Five things
no terminal ships.

Professor X

The lanes below are powerful individually. The job above them is to turn their evidence into one readable brief per ticker — what happened, what the demand says, what entered the radar and why.

That's Professor X: the brief-writing layer. Dossiers are drafted from every lane's evidence — filings, structure, OSINT, demand — in one voice, and every brief is graded afterwards against what the tape actually did. Not a chatbot; a desk editor with a scorecard.

OSINT radar

Geopolitical rhetoric. Maritime movement. Flight-pattern anomalies. Congressional disclosures from the primary record. SEC Form-4 insider prints as they file. Cyber-threat telemetry.

Fused, severity-weighted, and routed to the tickers they actually touch — from official and primary sources, on purpose. An intelligence lane, not a rumour mill.

• POLITY Head-of-state rhetoric · geopolitical posture
• CONGRESS Member trading disclosures · primary-sourced
• VESSEL Maritime movement & anomaly scoring
• FLIGHT Aircraft pattern tracking
• INSIDER SEC Form-4 open-market buys — disclosed facts
• CYBER Vulnerability & threat telemetry

The gates

Twelve thousand tickers in. A handful out. We studied 29,822 runner-days and learned the honest lesson: nothing picks winners on its own — discipline is refusing the losers. So every gate we ship is a suppressor with a documented reason to exist.

No blinking toasters. No FOMO noise. When Project Magneto surfaces a name, it's because the demand was real — absolute volume and float turnover, not a ratio that flatters thin tape — and because nothing in the situation tripped a rule the tape taught us the hard way.

The specimen library

When the tape prints something worth learning, we dissect it to the minute — entry, exit, failure mode, the exact mechanics — and the post-mortem becomes a screen the system runs live.

XPON's rename-pivot press release. AIXI's story-less forced parabola. The screen distilled from the AIXI study caught TNMG at +30% the next morning — it ran to +110% in fifteen minutes. The library grows every week, and the misses are in it too: the specimen we study hardest is always the one that got away.

The brains

A deterministic, sub-millisecond gate decides what deserves attention. A reasoning model writes the why — in plain language, constrained to a fixed vocabulary. The numbers are never the model's to invent: Python owns every figure, the model owns the prose.

Behind it sits a knowledge base built the slow way — market classics and ninety graded lessons of a top analyst's teaching, distilled into reference the briefs can cite. AI is leverage here, not an oracle: anything a model asserts about the tape must survive the grading loop like everything else.

The Referee

The lane that makes the others honest. An independent evaluation plane — its own database, its own identity, a sealed 400-day ledger no producer can edit — scores every signal source against the tape, on data it has never seen.

Its verdicts have teeth. When a quarter-million graded snapshots showed a six-voter ensemble of ours was inverted, all six were retired the same day. And no ranked call — no conviction band, no directional label — ships to a subscriber until the Referee certifies it on held-out tape. We'd rather show you the reason and stay silent on the rest than sell you a number we haven't earned.

The Community

Above the lanes is Professor X. Around them, we're seating a room: a small, vetted circle of professionals who read the tape for a living — being assembled now, desk by desk, from the first invites.

Members will annotate signals, flag false positives, and steer what gets built next. Their feedback shapes the system's rules and screens — through the same grading loop everything else must survive.