Build the rule. Prove it. Run on Autopilot.

A prediction market bot should show its work

Autopilot is built around a simple sequence: define the behavior, observe it on real market activity, and decide whether it deserves real capital. A paper strategy is not hidden behind a single performance number. You can open the rule and inspect the trades that produced its record.

The same workflow works whether your signal starts with a trader or a thesis. Follow-trader rules react to the entries of accounts you select. Strategy rules filter the market feed for the exact conditions you define. Both become automated trading bots only after you choose to promote them.

What happens in the seconds after a trade prints

A trade by an account you follow reaches Rivo through the live market feed, gets classified, and is matched against every Autopilot rule that could act on it. Kalshi entries run on a dedicated fast path with an age gate: if the trade is already older than the window you would have wanted to enter in, Autopilot skips it instead of chasing a price that has moved. Arriving late is treated as a reason not to trade, not something to paper over.

The same event can satisfy more than one of your rules. Each is evaluated on its own and each one that qualifies produces its own order, so a trader you follow and a strategy that happens to match the same market do not silently collapse into a single position.

Every skip is recorded, not just every fill

Most automation shows you only what it did. Rivo writes a row for every evaluation, including the ones where it declined to act, with the reason attached: the trade was too old, the daily limit was already reached, the account was not connected, the market was too thin to fill sensibly. When a rule underperforms you can tell the difference between one that picked badly and one that almost never fired.

Paper runs are recorded the same way, with the simulated fill price and size stored alongside the decision. A paper record is a log you can audit line by line, not a single number you are asked to trust.

Exits are rules too

An entry rule without an exit rule is half a strategy. Each position can carry a standing exit: a stop loss, a take profit, or a trailing stop that follows the high-water mark. Rivo marks open positions against the live bid roughly every 45 seconds and sells when one of your thresholds is crossed, so an exit does not wait on you being at the screen.

When Autopilot pauses itself

Three consecutive failed orders on a follow pause it automatically, and you get a notification naming the trader and why it stopped. Market conditions are deliberately excluded from that count. A book too thin to fill or a price that drifted past your limit is the rule correctly declining to trade, and counting those would pause your best-behaved follows first. Only genuine execution failures accumulate. You can also pause or resume any follow yourself at any time.

Orders that never fill

A resting order you believe filled is worse than no order at all. Every Autopilot order is reconciled against the venue itself: anything still unsettled after five minutes is checked directly with the exchange and marked filled, partially filled, or canceled, and an entry still resting after fifteen minutes is canceled outright rather than left to fill into a market that has already moved on.

Your account. Your keys. Your limits.

Rivo does not require a pooled wallet and never takes custody of your funds. Live Kalshi execution runs through your own Kalshi API key, encrypted at rest and used for nothing except placing the orders your rules produce. Disconnecting the account wipes the stored key.

You set the amount per trade and the maximum per day, per follow and across the account. The daily ceiling fails closed: if the spend so far cannot be confirmed, the order does not go out. Live trading also requires an explicit confirmation beyond simply enabling a rule, so nothing moves real money because a default was left in place.

Autopilot opens matching entries and logs what it did. It does not promise profit, and paper results do not guarantee live results.

Live on Kalshi today. Polymarket coming soon.

You can run real-money Autopilot on Kalshi now. Polymarket traders and markets are available for research, alerts, backtesting, and paper strategies, but live Polymarket execution is not enabled yet. For the execution details, read about the Kalshi trading bot. For the broader bot workflow, see our prediction market trading bot guide.

Frequently asked questions