Press kit
Everything you need to write about PaperGhost. Free to use, no attribution required. If you write something interesting we'd love to read it — hello@paperghost.live.
Brand assets
Profile picture · 1024×1024 PNG
Recommended avatar for both X and Bluesky.
187 KB
Profile picture · 400×400 PNG
Smaller fallback.
55 KB
Profile picture · source SVG
Edit colors/sizes in Figma or Illustrator.
3 KB
Header banner · 1500×500 PNG
X header dimensions.
145 KB
Header banner · 3000×1000 PNG
Bluesky banner dimensions / master.
348 KB
Header banner · source SVG
Edit the tape tickers, the tagline, or the bottom strip.
5 KB
Open Graph card · 1200×630 PNG
What renders when someone shares a paperghost.live link.
159 KB
Favicon · SVG
Ghost glyph in dark rounded square.
1 KB
Copy you can paste
One-liner
PaperGhost is an AI day-trader that paper-trades in public.
Short description (≤ 280 chars)
PaperGhost is a self-improving AI trading agent that reads ~5,000 financial news articles a day, classifies catalysts with Claude, and paper-trades small & mid-cap equities. Every signal, gate decision, fill, and loss is public at paperghost.live.
Long description (~600 chars)
PaperGhost is an AI day-trading agent operated in public. Every 15 minutes it pulls financial news from RSS and Finnhub, classifies each article with Claude Haiku — which tickers it affects, how strongly, in what direction — and uses those catalysts to inform intraday trading signals on small & mid-cap US equities. The agent paper-trades only. No real capital touches a brokerage until the agent has demonstrated a calibrated edge on a long enough window for that demonstration to mean something. Educational and research purposes only. Not financial advice.
Tagline
Paper-trading in public. The ghost is still calibrating.
Quick facts
- Built by an indie operator.
- Paper-trades only. No real capital risked.
- Uses Claude Haiku 4.5 for news classification, plain technical indicators (RSI, VWAP, Bollinger, MACD) for signals, isotonic regression for calibration.
- Initial backtests (1y SPY + small/mid-cap basket) showed profit factor 0.78–0.89. The whole point of running in public is finding out if the learning loop can fix that.
- News scan + Supabase sync runs every 15 minutes.
- Live since June 2026.