LIVE
MARKETS
📁 PORTFOLIO
—
click any row to load chart
Cumulative P&L Curve
Analytics — Moonshots
Sector Allocation
60-Day Return Correlation
Earnings & Price Alerts
Position Performance
Risk Metrics
🎓 Real Money Graduation Tracker
WudiRL Quant Terminal
AUTONOMOUS AI PAPER-TRADING ENGINE · v8.0
TWO PORTFOLIOS, ONE ENGINE
🚀 MOONSHOTS
High-volatility momentum & breakout stocks $3–$50.
Runs every trading day (Mon–Fri). Up to 5 picks/day.
Conviction sizing: $150 / $250 / $350 per position.
Minimum conviction: ★★ medium or higher.
Exit rules: +20% target / -10% trailing stop.
📁 CORE LONG-TERM
Quality mid-to-large cap growth stocks $5–$200/share.
Runs every trading day. Max 10 concurrent positions.
Conviction sizing: $200 / $300 / $400 per position.
Minimum conviction: ★★ medium or higher.
Exit rules: +40% target / -15% trailing stop.
HOW IT WORKS
8:30AM
AI Research Run (Moonshots) — Claude Sonnet 4.6 reviews all past trades and ML-derived win rates by sector, then picks up to 5 momentum/breakout stocks ($3–$50). Only ★★+ picks deployed. Conviction sizing: ★★=$250 / ★★★=$350. Budget: $1000/day.
8:30AM
AI Research Run (Core Long-Term) — Claude picks quality mid-to-large cap growth stocks ($5–$200) with strong secular tailwinds. Diversifies across sectors. Max 10 concurrent positions. Conviction sizing: ★★=$300 / ★★★=$400. Budget: $1000/day.
WEEKLY
Re-Conviction Review — Open moonshot positions held 10+ days and in the float zone (-15% to +15%) are re-analysed by Claude. SELL verdict on a winning position (+3%) overrides to HOLD. SELL on a flat/losing position held 14+ days closes it immediately.
DAILY
Portfolio Audit (Both) — Each position evaluated against its exit rules. Moonshots: +20% target / -10% trailing stop. Long-term: +40% target / -15% trailing stop. Green candle days automatically ratchet the trailing stop up to lock in gains.
DAILY
Dashboard Rebuild — Deploys to GitHub Pages with fresh data. Finnhub live prices update every 60s during market hours (9:30am–4pm ET) via the browser — no server required.
🧠 THE LEARNING ENGINE

Before every research run, Claude reads the complete history of closed trades — wins and losses alike. It extracts patterns: what thesis types worked, which sectors failed, what price levels got stopped out. Each new batch of picks is explicitly shaped by this analysis.

The re-conviction engine adds a second layer — open positions aren't just held until a stop or target fires. Claude actively revisits the original thesis after 10+ days and decides whether to stay in or cut. This prevents dead-money positions from tying up capital when the original catalyst has faded.

STRATEGY RULES AT A GLANCE
🚀 MOONSHOTS
📁 CORE LONG-TERM
${[ ['Universe', '$3–$50 US stocks', '$5–$200 US stocks'], ['Frequency', 'Daily (Mon–Fri)', 'Daily (Mon–Fri)'], ['Max Picks', '5 per day', '10 concurrent max'], ['Min Conviction', '★★ medium+', '★★ medium+'], ['Position Size', '$150 / $250 / $350', '$200 / $300 / $400'], ['Target', '+20%', '+40%'], ['Stop Loss', '-10% trailing', '-15% trailing'], ['Re-conviction', 'Weekly (10+ days held)', 'At target/stop only'], ['Daily Budget', '$1,000 / day', '$1,000 / day'], ].map(([label, ms, lt], i) => `
${label}
${ms}
${lt}
`).join('')}
🔔 PUSH NOTIFICATIONS

Native push notifications fire directly to your device — no email, no Telegram, no third-party service. The service worker checks for new events every 5 minutes and fires alerts for: position closed (with P&L), near target, near stop, or daily heartbeat. Works for both portfolios.

Checking notification status...
Built on GitHub Actions · Claude Sonnet 4.6 · Finnhub · yfinance · Zero cost infrastructure
For educational and paper-trading purposes only. Not financial advice.
Loading backtest data...