Last updated: 2026-04-18 · Plain English, not legalese
We collect data that helps the product. We never collect data that puts you at risk.
Concretely: your Kalshi credentials never touch our servers (we can't steal them, we can't subpoena them, we can't sell them). Your trade history and match queue ARE stored, because the dashboard is useless without them and we need them to calculate your referral commissions.
When your bot trades, it sends us the prediction it made (e.g. "model says 73% KKR wins at over 12") plus the outcome when the match ends. Aggregated across all users, that's how we know if the model is calibrated or drifting.
Three things we use aggregate data for:
None of this links back to you personally. A row in our calibration table is just: (prediction %, actual outcome, match total, venue, model version). No user ID.
YOUR LAPTOP OUR SERVERS
----------- ------------
kalshi private key ❌ NEVER (not transmitted)
config defaults ❌ NEVER
local trade log mirror → trades table (dashboard)
match queue mirror → match_queue (dashboard)
user account + billing
referral graph
KALSHI (third party)
--------------------
authoritative orders + positions
actual account balanceNo, except with Stripe (for payment processing) and Kalshi (which you authorize when you set up the bot). We do not sell, rent, or barter user data. If we were ever acquired, any new owner would be bound by this policy as of the date of acquisition.
We'd rather you ask than guess. help@freshloop.ai