A live view of an AI multi-agent trading pipeline. Each pixel character on
the floor is one agent — news, research, sentiment, time-series,
strategy, mixer, judge, and so on. Their state updates as the daily
pipeline runs on the host machine.
HOW TO READ THE FLOOR
working / done — agent active or finished
alert / error — flagged signal or failure
The status indicator (top-right) reads LIVE when the pipeline is running, IDLE between runs, OFFLINE if the host is unreachable
Status refreshes every ~10 seconds
WHERE TO FIND THINGS
Reports tab (right panel) — daily PDF picks per asset class, kept for 30 days
Agent / History / Picks / Charts / Pipeline tabs — drill into per-agent activity and prior runs
SUBSCRIBE (top-right) — one email per run with that day's picks
🐛 (top-right) — report a bug
Educational project for Pace University CS668. Not investment advice. PDFs auto-expire after 30 days.
Edit Tickers
X
Changes apply to the next pipeline run. Tickers are saved in memory only.
🧠 NEURAL MAP
X
Trading memory graph — nodes sized by importance, colored by asset class, connected by relationships
★ POSITIONS
X
Open and recent positions with stop-loss / take-profit tracking
🐛 BUG LEDGER
X
Known issues tracked in DuckDB — persists across reboots