She sees it before
the market does.
A Cassandra is an AI that monitors one specific corner of the world — SEC filings, jobs reports, FDA recalls, power grid stress — and trades your brokerage account when it spots something. Every action comes with a plain-English explanation of what it saw and why it acted.
Each Cassandra watches one thing. All day. Every day.
Every Cassandra is a specialist. She monitors specific government and regulatory data sources, and when something happens that moves markets, it tells you what it found and what it wants to do about it.
Cassandra: Dilution
Wave 1When a company quietly files paperwork to sell more shares, existing shareholders get diluted. This Cassandra reads every SEC filing as it hits EDGAR and spots the capital raises, shelf registrations, and emergency financing that usually tank a stock before most people notice.
"ACME Corp just expanded its shelf registration. Financing language suggests they need cash soon. Previous shelf was 80% used. Shorting ACME."
Cassandra: Filing Stress
Wave 1Buried in every quarterly and annual report is language that signals trouble — going-concern warnings, covenant breaches, auditor red flags, deteriorating customer concentration. This Cassandra reads every 10-Q, 10-K, and 8-K looking for the sentences that say a company is in trouble before the headline writers catch on.
"Going-concern language just appeared in XYZ's 10-K that wasn't in last quarter's filing. Auditor is flagging liquidity risk. Shorting XYZ."
Cassandra: Macro
Wave 1CPI, jobs numbers, GDP, retail sales, Fed announcements — these reports move the entire market in seconds. This Cassandra knows the release calendar, knows what Wall Street expects, and trades the surprise. Not the number itself, but how far it is from what everyone was betting on.
"CPI came in at +0.4% vs the +0.2% everyone expected. Inflation is re-accelerating. Bond prices are about to drop. Shorting TLT."
Cassandra: Grid Stress
Wave 1Extreme heat, polar vortexes, and storm systems stress regional power grids. When ERCOT or PJM start issuing warnings, energy stocks move. This Cassandra monitors National Weather Service alerts and EIA grid data, and trades utilities and energy companies when weather and grid stress line up.
"Heat dome parked over Texas and Oklahoma. ERCOT demand forecast is spiking. Grid stress probability is elevated. Going long NRG."
Cassandra: FDA
Wave 2Drug recalls, manufacturing shutdowns, shortage alerts, rejection letters — the FDA publishes all of it. This Cassandra watches for the operational and quality problems that crush pharma and device stocks, not the binary approval-or-not gambles.
Cassandra: Policy
Wave 2Tariffs, antitrust actions, drug pricing rules, AI regulation, sector-specific rulemaking — it all gets published in the Federal Register and Congress.gov before it hits the news. This Cassandra reads the firehose of government policy so you don't have to.
Cassandra: Agriculture
Wave 2USDA crop reports and supply estimates move commodity markets the instant they drop. These numbers are highly structured, released on a known schedule, and historically huge market movers — but almost no one is watching them systematically.
Pick your Cassandras. Set your rules. Let them run.
Choose What to Watch
Browse the lineup and pick the Cassandras that match what you care about. SEC filings, economic data, weather events, FDA actions, government policy — each one covers a specific domain.
Decide How Much to Risk
Allocate capital to each Cassandra separately. Set position limits, stop losses, and how aggressive you want it to be. Every Cassandra manages its own budget independently.
Choose Your Level of Control
Three modes: Alert (just notify me), Propose (show me the trade and let me approve it), or Auto-Execute (just do it). You can change this anytime, per Cassandra. Start cautious. Turn up the dial when you're ready.
Watch It Work
Every action comes with a plain-English explanation: here's what I saw, here's the source, here's why I'm acting. Track performance per Cassandra. See exactly what it read and why it traded.
Every trade shows its work.
No black boxes. When a Cassandra acts, you see the original source, what it concluded, and exactly why it's making this trade right now.
SEC filing detected: ACME Corp (ACME) filed an S-3 shelf registration amendment. The financing language points to a near-term capital raise. Their previous shelf was already 80% used up.
High probability they issue new shares within 30 days. The language looks like emergency financing, not routine. It's a small-cap with few institutional holders — that means the stock price will likely take a bigger hit than usual.
Bureau of Labor Statistics just released CPI: +0.4% month-over-month vs. the +0.2% everyone expected. Core services ex-shelter came in at +0.5% — the hottest reading in four months.
Big upside surprise. Inflation is re-accelerating in the stickiest categories. The market had been pricing in rate cuts — that's about to get unwound. Long-duration bonds are vulnerable.
This isn't a research tool. It actually trades.
One Cassandra, One Job
Each Cassandra watches a narrow slice of the world — not "AI reads the entire internet." Dilution watches SEC filings. Macro watches economic releases. Grid Stress watches weather and power data. Focused beats general.
Trades Your Actual Account
Connect your brokerage account. When a Cassandra triggers, it executes for real. Your money stays at your broker — we never touch it. Currently supports Interactive Brokers and Tradier.
You Control the Dial
Alert, Propose, or Auto-Execute — per Cassandra, changeable anytime. Start by just getting notifications. Move to approving each trade yourself. Flip to fully automatic when you trust it.
Split Your Capital
Run multiple Cassandras at once, each with its own budget. Put $30K behind Macro, $20K behind Dilution, $10K behind Grid Stress. Each one manages its own allocation and risk independently.
Plain-English Explanations
Every trade comes with a "here's what I saw and why I acted" in words you can actually read. No mystery signals. No "trust the model." You see the source document and the reasoning.
Track Each One Separately
See P&L per Cassandra. Know exactly which one is making you money and which one isn't. No blended numbers. If one isn't working, turn it off. If one's crushing it, give it more capital.
Government Data, Not Guesswork
Every Cassandra watches official public sources — SEC EDGAR, BLS, the National Weather Service, openFDA, the Federal Register. Real data from real agencies. You can verify everything it sees.
Risk Controls on Everything
Position limits, stop losses, allocation caps, drawdown thresholds. Every Cassandra checks your risk rules before every single trade. You set the guardrails.
Be first to deploy.
We're opening access soon. Leave your email and we'll let you know when the first Cassandras are live.
Questions.
Important Disclosures
Not financial advice. The Cassandra Project provides software tools that enable automated event-based trading based on parameters defined by the user. We do not provide investment advice, portfolio management, financial planning, or recommendations to buy or sell any securities.
Self-directed accounts only. All trading occurs in the user's own self-directed brokerage account, maintained at a third-party broker-dealer of the user's choosing. Users are solely responsible for all investment decisions and for the activity in their brokerage accounts.
Risk of loss. Trading securities involves substantial risk of loss. Automated trading systems carry additional risks including technology failures, delayed event detection, incorrect signal interpretation, market gaps, and execution slippage. Past performance is not indicative of future results. You could lose some or all of your invested capital.
AI limitations. Event detection relies on publicly available data sources and natural language processing, which may be delayed, incomplete, or inaccurate. Cassandras may misinterpret events, generate false positives, or fail to detect relevant events entirely. Users should not rely solely on automated systems for investment decisions.
Hypothetical performance. Any performance data, examples, or simulations shown are hypothetical and do not represent actual trading results. Hypothetical performance has many inherent limitations, including the benefit of hindsight.