We don't predict.
We recognise.
Project Magneto sweeps twelve thousand tickers every session — the filings, the wires, and an OSINT lane no terminal ships — and surfaces the handful of situations that survive its gates, each one carrying the reason it entered the radar. Then it grades itself on what happened next.
- Tickers swept
- 12,000+
- every session · pre · RTH · AH
- Autonomous agents
- 140+
- specialised · coordinated
- Graded snapshots
- ¼M +
- outcome-scored vs the tape
- Misses published
- Yes
- every one becomes a rule
Real demand, named. Manufactured moves, called.
Absolute volume and float turnover decide what's real — not hype, not a headline. When a move is mechanically forced, we say so. When demand is genuine, we say that too.
From the lab notebook
the archive · opening soon →Real, dated entries from the system's own record — detections and misses. Percentages are what the public tape did, shown as observational record — not trading results, and not a promise.
A scrolling record of dated system detections and misses: AIXI, XPON, TNMG, JUNS caught and graded; ZSTK and SLV are documented misses that led to rebuilt lanes. Observational record, not trading results.
We built this for ourselves.
Now we're letting you in.
Project Magneto began as a private trader's war room. Every module exists because we needed it at 09:29 ET. Nothing shipped to sell — only to survive the tape.
We grade ourselves.
We retire what fails.
Every alert is scored against what the tape actually did. When a quarter-million graded snapshots showed one of our own signal ensembles had no lift, we deleted it — in one day. That bar never moves.
The reason is the product —
not by AI. With a lot of help from it.
Every alert carries why it entered the radar, in plain language — filings, structure, OSINT, demand — written to be checked, not believed. AI is leverage. Judgement stays yours.
Peer-led pipeline.
Tell us what you need.
Wall Street tools are closed boxes. Project Magneto isn't. The pipeline is ours end-to-end, and member requests drive what gets built next — if it's useful to one serious trader, it's usually useful to the room.
Nowadays, anyone can build a trading bot in a weekend. Buy an API key, wrap a model, ship a webhook.
We're not a trading bot.
We're a decision-making ecosystem.
140+ coordinated agents. An OSINT lane the street doesn't ship. A grading loop that scores every alert against the tape — and an independent Referee that must certify any ranked call on held-out data before it ever reaches a subscriber. The button stays yours; we're the room that shows you what the tape is doing, and how sure we've earned the right to be.
Meet Professor X.
The one who connects them all.
The one who sees them all.
A hundred and forty specialised agents are only as valuable as the layer that turns their evidence into something you can read. Professor X is that layer — the brief-writer that draws on every lane and hands you one dossier, in one voice, for the ticker in front of you. And every brief it writes is graded afterwards against what the tape actually did.
- READSevery lane, every session
- WRITESone brief, one voice, per ticker
- CITESfilings · structure · OSINT · demand
- GRADEDagainst the tape, after the fact
Not a chatbot. Not a summariser. A desk editor with a scorecard. The button stays yours — Professor X writes the brief; the grading loop decides whether it was any good.
Not just a product.
A room.
A vetted circle of professionals.
Being seated now — desk by desk.
Project Magneto is assembling a room around the system: traders, quants, discretionary PMs, options specialists, macro readers, sector operators — connected to each other and to the tape. You don't buy a seat. You earn one.
You'll share with us.
We'll share with you.
Contributions raise the ceiling for everyone. Annotations, false-positive flags, feature requests — the room remembers, and the room gives back.
Terminals are one-way. Public forums are noise. This room is the middle — quiet, sharp, signal-only, and intentionally unseen.
- MACROUS / DM rates
- OPTIONSvolatility surface
- DISCRETIONARYUS small-caps
- STRUCTURALorder-flow · intraday
- SECTORsemiconductors
- EVENT-DRIVENM&A · spin-offs
- QUANTsystematic · execution
- CREDITIG · HY
- ENERGYcrude · refined
- MICROSTRUCTUREliquidity · tape reading
- GEOPOLITICSregional risk
- GENERALISTfull-time tape
What separates us
the five lanes →An intelligence lane no terminal ships
Geopolitical rhetoric, maritime & aircraft movement, congressional disclosures from the primary record, SEC Form-4 insider prints as they file, cyber-threat telemetry — fused, severity-weighted, routed to the tickers they touch.
A scanner that suppresses, honestly
Twelve thousand tickers in, a handful out. Freshness, reversal, real-demand and liquidity gates — each one exists because the tape taught us a specific, documented failure. We publish what the gates cost, not just what they catch.
A library of setups, studied to the minute
XPON's rename-pivot. AIXI's forced parabola. Each specimen is dissected — entry, exit, failure — and becomes a screen the system runs live. The screen born from AIXI caught TNMG at +30% the very next morning.
Deterministic gates. A model that explains.
A sub-millisecond probability gate decides what's worth attention; a reasoning model writes the why in plain language, constrained to a fixed vocabulary. The numbers are never the model's to invent.
An independent judge with a locked ledger
A separate evaluation plane — its own database, its own identity, a sealed 400-day record — scores every producer. No ranked call ships to a subscriber until the Referee certifies it on data it has never seen.
Verification receipts. Named failures. No band-aids.
Every miss gets a dated post-mortem and becomes a rule. When six of our own signal voters graded inverted, we retired all six the same day. Reality grades us — and we keep the grade.
If the tape is your job,
Project Magneto is your edge.
Tell us what you trade. If you're a fit, we'll open the door.