I built an AI that reads 1,353 financial articles and only trades on 4.6% of them
Most trading signals are backwards — they start with a chart pattern and rationalize a story around it.
I wanted to flip that. Start with the catalyst. Verify the edge. Then find the contract.
So I built Oracle. It monitors thousands of financial news articles, SEC filings, and FinTwit accounts every 60 seconds. When it detects one of three specific catalyst types — earnings beats with raised guidance, phase 2/3 clinical trial results, or credible fraud reports — it fires a complete options trade setup.
Every signal includes the exact contract, live bid/ask/mid pricing, full Greeks (IV, Delta, Theta, Gamma, Vega), open interest, volume, and a confidence rating.
The backtest: 365 days. 144 tickers. 1,353 articles. Blind classifier — the model never saw the outcomes during training. 64.3% directional accuracy at 48 hours.
The selectivity is what matters most. Oracle reads everything but fires on 4.6% of articles. 1-3 signals per day, not a firehose.
This isn't financial advice — it's an AI analysis tool for options traders who want a systematic edge rooted in real catalysts, not vibes.
$150/month. Cancel anytime.
