Kalshi trading bot

The engine is crypto-first on Kalshi. It scans the configured series, by default the 15-minute and hourly BTC and ETH markets, picks the most active open market, classifies it with the same taxonomy used on every venue, and flies all nine crypto strategies against it with the full signal stack. As each market closes the rotation loop moves to the next one, so there is always a live squadron on the asset.

Sports and politics markets are deployed from the Control Tower's Squadron Builder. A sports squadron matches a Kalshi NFL or MLB game market to the bookmaker event and rests bids under the vig-free consensus line. A politics squadron runs the maker and arbitrage strategies on a long-dated market. The Helm lets you enter any single Kalshi market with a thesis, a probability and a falsification condition.

Fees. Kalshi charges a taker fee that scales with price and peaks at 1.75 cents per contract at 50 cents. DRADIS prices that schedule into each strategy's edge threshold, so the maker-first strategies lean on resting orders and the taker strategies demand enough edge to clear the fee.

Order books. Books stream over Kalshi's authenticated WebSocket as deltas with sequence numbers. The engine folds each delta into its local book and resynchronizes from a snapshot when it detects a gap, so a strategy never quotes against a stale book.

Fills. Fills arrive on the private fill channel and are reconciled against the engine's own position book event by event. A position that exists on Kalshi but not in the database is adopted with the right label rather than ignored.

Price history. Kalshi does not publish a public price history for a market. The Control Tower's Markets page says so plainly rather than drawing an empty chart, and shows the depth and recent tape it can read.

Kalshi's crypto contracts settle on the underlying, so the engine watches the underlying directly. The Price Raptor streams Binance spot with velocity and acceleration, the Funding and Derivatives Raptors add the perpetuals' funding rate, open interest and cumulative volume delta, and during US cash hours the Tide Raptor reads ETF flow from Alpaca IEX. Each strategy reads those feeds and the Kalshi book every tick. FairValue computes a fair probability from realized volatility, distance to strike and time to close and buys only below it by your margin. Momentum fires on a Binance move the Kalshi book has not absorbed. Maker quotes both sides and manages net exposure.

Two paths. The Kalshi demo exchange at demo.kalshi.co gives you mock funds and its own API keys, and DRADIS trades it through exactly the production code path. Ghost mode goes the other way: it reads the real production books and simulates the fills, so you can measure a configuration against real liquidity without an order ever reaching the exchange.

The Control Tower is a web dashboard served by the instance itself. It shows P&L, open positions, the trade log and every strategy's current verdict with the gate that is holding it, under the Kalshi asset selector. Live configuration is edited there with schema-validated bounds. An optional AI advisor, local or hosted, reads the trade history and proposes configuration patches under an autonomy tier you choose, from recommend-only to fully autonomous with a circuit breaker.

Install and upgrade guides are at dradis.live/docs. The same engine also runs on Polymarket.