Fin45's AI agent ingests 50+ data feeds across 10 signal categories. Every trade decision starts here.
SEC Form 4 filings from corporate officers, directors, and 10% shareholders.
Financial disclosures from members of Congress and their spouses.
Unusual options activity — large sweeps, block trades, and high OI changes.
Off-exchange volume anomalies and institutional block prints.
13F, 13D/G, 8-K, and other material filing events from EDGAR.
Pre and post-earnings whisper numbers, guidance shifts, and revisions.
Fed minutes, CPI/PPI data, yield curve shifts, and economic indicators.
Academic papers, patent filings, and R&D activity signals.
Major litigation, SEC enforcement actions, and regulatory decisions.
Polymarket, Kalshi, and Metaculus probability shifts on market-moving events.
Raw data from SEC EDGAR, congressional disclosures, options chains, dark pools, court filings, and more.
Each signal scored 0–1 on conviction. Only signals ≥ 0.75 advance to trade consideration.
Multi-signal confluence required. Position sizing from Kelly criterion + volatility adjustment.
Paper trades placed via Alpaca. Entry, target, and stop published 10 days after close.
Fin45 ingests 50+ feeds including SEC EDGAR (Forms 4, 13F, 13D/G, 8-K), congressional financial disclosures, options flow from OPRA, dark pool prints, earnings transcripts, Fed minutes, academic research from arXiv, patent filings from USPTO, court dockets from PACER, and prediction market probabilities.
A signal must score ≥ 0.75 conviction (on a 0–1 scale) to qualify for trade consideration. The agent then requires multi-signal confluence — a single data point is never enough. Position sizing uses a modified Kelly criterion adjusted for realized volatility.
Fin45 operates a 10-day embargo. Trades are revealed only after the position has been closed for 10+ days. This prevents front-running and ensures no one can copy trades in real time.