Your always-on Kalshi trading assistant—built for short-duration markets. It scans, ranks under strict rules, sizes responsibly, and only fires orders when setups clear every gate you configured.
Runs on repeating cycles so you never miss nearby resolutions. Targets near-close markets, ranks them by confidence and edge, optionally blends AI + heuristics + flow signals, allocates within bankroll constraints, and skips a cycle entirely when nothing is safe enough.
Built-in brakes: daily loss vs start-of-day cash, correlated event caps, cooldowns between related trades, losing-streak pauses, shutdown after consecutive failures, optional hourly goal overlays, telemetry for forensics—and a GUI to steer it quietly.
This is not “always be trading.” It’s always be ready—with the discipline encoded so you can’t impulse-click your way into the wrong side of a wide book during a chaotic hour.
Exactly what happens when the runner wakes up—from cash check to telemetry write.
1) Refresh settings and account state: Reloads config snapshot, grabs balances—or paper wallet in dry run—updates dashboards, rotates daily anchors when your calendar flips.
2) Apply safety rails first: Daily loss lockout, enforced-goal / high-risk flags, streak pauses, blocked local hours—all before scanners hit the chain.
3) Scan and filter: Pull near-resolution markets inside your horizon, shred anything failing spread, liquidity, structural gates (bundle integrity, deadline caps).
4) Decide: Rank with your chosen intelligence stack, reconcile optional consensus + cross venue + leaderboard-flow nudges, choose top-lineups with correlation + cooldown logic.
5) Size and place: Convert picks into order intents via Kelly or split rules, enforce daily trade caps, refresh order books on finalists, submit live or simulate with paper cash.
6) Learn and log: Telemetry rows for predictions and orders, adaptation pressure from P&L, calibration + auto-tune hooks, recent bet settlement refresh—then sleep until the next interval.
Every tag below maps to real controls in config.yaml and the desktop settings panes.
Dial in conservative, balanced, or aggressive personalities—the bot replays that profile every loop without fatigue.
The feature page groups every capability by subsystem—runtime, strategy, risk, AI, gates, sizing, and more.
