DeskBot for Kalshi vs Poly Sniper

What DeskBot for Kalshi does

Desktop bot for Kalshi's 15-minute up-or-down markets, all 19 of them at once: nine coins — BTC, ETH, XRP, SOL, DOGE, HYPE, ZEC, NEAR and BNB — plus gold, silver, copper, platinum, palladium, oil, natural gas and three currency pairs. It runs on your own Mac or Windows PC, trades your own Kalshi account through your API key and writes a reason for every trade and every skip on a local dashboard. You choose the markets, the contract size, which side it may take and a cap on what it can risk; it weighs each price against the time left and against the win rate that price needs after Kalshi's fee. In the developer's own live account it buys the side already priced around 92¢, so a win pays about $1 and a loss costs about $14. $19.99 a month, with a free 24-hour trial in paper mode. Kalshi only — it does not trade Polymarket.

• All 19 of Kalshi's 15-minute series on one screen, each switched on or off by itself, with its price, the time left and the bot's call
• A written reason for every trade and every skip — “too close”, “near certain · fee too big”, “watching up side” — on a dashboard served by the app on your own computer
• Your rules: contract size, the yes side, the no side or both, and a cap on what it can risk
• A break-even check before each entry: the win rate the price needs after Kalshi's fee, against the bar you set
• Stays out of the last stretch of a window rather than open a trade with no time left to be right

• The developer's own live account shows what the bot does by default. From Oct 8 to Oct 11, 2026 it made 336 trades, won 317 and lost 19, and netted +$64.06 after Kalshi's fees, buying the side already priced at 85–95¢ — 92.1¢ on average — with 4 to 11 minutes left in the window. At that price a win on a 15-contract order pays about $1 and a loss costs about $14, so the bot has to win roughly 92 times in 100 just to break even, and one loss erases a dozen wins or more. The first day shows the risk: 108 wins, 12 losses and −$51.64. The site says all of this itself, plainly, and calls it one account over a short stretch. Its results page has not been updated since Oct 11; a frame of the vendor's demo from Oct 14 reads +$140.05 over 712 settled trades, and −$41.39 for that day.
• We checked that ledger against Kalshi rather than taking it on trust. For 35 trades — the 25 on the results page and 10 more shown on the developer's Kick stream on Oct 12, EUR/USD and oil among them — the side Kalshi settled each window on matched the win or loss recorded, and Kalshi's public trade tape shows fills on the recorded side at the recorded price in that minute; in the EUR/USD window the only such fill is a single 15-contract order at 93¢, the size the ledger implies. Every profit and loss we tried recomputes to the cent from Kalshi's fee formula. The demo video on the home page is the working app, not a mock-up: the strikes it shows for the 3:15 PM ET window on Oct 14 — BTC 81,629.01, ETH 2,429.42, XRP 1.3279 — are the ones Kalshi set. None of this proves whose account it is, or that your copy will behave the same: the video shows the developer's own build, and the vendor says your results will differ with your settings, size and timing.
• Your money stays in your Kalshi account, and by the vendor's description your API key stays on your computer: the app asks for it when it first starts and serves its dashboard locally. A Kalshi API key can be revoked on your own account page, so create one for DeskBot alone. The trust question moves to the program itself. The builds are not signed — on a Mac you open the app with right-click → Open, on Windows past SmartScreen with More info → Run anyway — and what you run is closed code from a vendor whose domain was registered on Sep 20, 2026.