Features
ScanForge is built on a single source of truth: one analysis engine, one scoring pipeline, consumed by CLI, web UI, and REST API alike.
Claude AI Does the Reading for You
Every ranked trade gets a fundamental deep-dive: business model, earnings context, analyst view, key risks, and price targets — generated by Claude and delivered in plain English.
- ✓ Deep fundamental analysis per ticker
- ✓ Risk factors and red flags surfaced automatically
- ✓ AI re-ranker combines quant + news into final priority
- ✓ Macro tailwind / headwind detection prevents bad reversal calls
News That Actually Moves the Market
Finnhub is the primary news source; Grok / xAI is the automatic fallback. Every ranked ticker gets today's catalyst sentiment attached, and the AI re-ranker uses that context to promote setups with real momentum.
Finnhub Primary
Company news, earnings, filings, analyst changes — pulled per ticker on every run.
Grok / xAI Fallback
Automatic failover keeps the pipeline running when Finnhub rate-limits or goes down.
AI Re-Ranker
Blends news sentiment weight with quant score into a final priority list you can actually trade.
A Full $200K Virtual Account
Test every idea before you risk real capital. Automated monitoring polls prices every 5–10 minutes during market hours, emails you on stop loss / profit target hits, and warns you when options are about to expire.
Automated Monitoring
5–10 min price polling 9:30am–4pm ET. Daily summary at 4:05pm. Email alerts on every trigger.
Real / Virtual Tracking
Flag trades as real or virtual in a single account. Test strategies alongside your live book.
Trade Journal
Structured entries for every closed position: what worked, what failed, lessons. Export to markdown.
CSV Export
Tax-ready exports with P&L and holding periods. Date filters. One-click for closed / open / full log.
Know Your Exposure at a Glance
Aggregated delta, theta, gamma, and vega across every open position. Risk warnings fire when your book gets too concentrated in one direction.
Validate Before You Trade
34,000+ historical samples from Polygon.io. Walk-forward out-of-sample testing, Monte Carlo simulation, and win-rate validation on every signal you build.
Build Your Own Edge
50+ metrics available. Combine them into custom scan conditions, save as templates, and run across the full universe with one command.
Built-in signal types: CAT_PRESENT, DEAD_CAT_BOUNCE, reversals, momentum breakouts, volatility squeezes, and more. Or build your own from scratch.
Three Ways In, One Engine
Flask web UI on :5100, FastAPI REST on :8000/docs, and a full CLI — all backed by the same single-source-of-truth analysis engine. 13 paper trading endpoints. Kelly-Criterion position sizer. Automated email reports.
Flask Dashboard
Interactive analysis, position management, portfolio view — in the browser.
FastAPI REST
13 paper-trading endpoints plus analysis, backtest, and file routes. Full OpenAPI docs.
CLI Scripts
One-shot analysis, backtests, trade journaling, and CSV export from the terminal.