- A Kalshi bot strategy is a written rule a program can follow without you. The five common families are rule-based execution, providing liquidity, reference models, cross-venue price gaps and scheduled-event trading.
- Fees decide most of them. A bot that enters and exits 100 contracts at 50¢ as a taker pays $3.50 in fees, so the price has to move more than 3.5¢ a contract before the round trip earns anything.
- Kalshi's Developer Agreement limits API use to a member's own trading, lets Kalshi limit API calls at its sole discretion without notice, and prohibits facilitating trading by other members.
- A bot executes a rule; it does not make the rule right. Research on Polymarket accounts found only 3.14% were statistically skilled.
Disclosure: PredictionsMarket.com earns referral commissions from Kalshi when readers sign up through our links (see our affiliate disclosure). Nothing here is financial advice. Automated trading can lose money faster than trading by hand.
The right Kalshi bot strategy depends on three things you can check before writing any code: which side of the fee schedule the bot sits on, what outside data it needs, and whether Kalshi's API terms allow the way you plan to run it. This guide compares the five approaches on those three points. The fee figures use Kalshi's fee schedule effective July 7, 2026, and the terms come from version 1.1 of the Kalshi Developer Agreement, read on November 10, 2026, along with Kalshi's API documentation and help center (see sources). We do not run a trading bot, and no strategy below has been tested with money by us.
What a Kalshi bot strategy has to get right first
Kalshi contracts pay $1 if the outcome happens and nothing if it does not, so a price in cents reads as a rough probability. A bot can place two kinds of order. A taker order trades at once against an order already in the book. A maker order rests in the book until someone else trades against it. Our guide to how Kalshi prices and orders work covers the order book in detail.
The fee schedule treats the two differently. The taker formula is 0.07 × contracts × price × (1 − price), rounded up to the next cent. By default a resting order on Kalshi pays no fee; on the series where a maker fee applies, the rate is 0.0175 in place of 0.07. Because the formula peaks at 50¢, the fee is largest exactly where prices are most uncertain.
| Price | Fee to enter | Fee to enter and exit at the same price | Price move needed to cover fees |
|---|---|---|---|
| 10¢ | $0.63 | $1.26 | 1.26¢ |
| 25¢ | $1.32 | $2.64 | 2.64¢ |
| 50¢ | $1.75 | $3.50 | 3.5¢ |
| 75¢ | $1.32 | $2.64 | 2.64¢ |
| 90¢ | $0.63 | $1.26 | 1.26¢ |
A position held to settlement pays the entry fee only. A bot that trades in and out many times a day pays the round trip every time, and it also pays the spread, the gap between the best bid and the best offer. Run your own numbers in the fee calculator, or read the full worked examples in our break-even guide to prediction market fees.
Five Kalshi bot strategies compared
| Strategy | What the bot does | Fee side | Outside data | Main way it fails |
|---|---|---|---|---|
| Rule-based execution | Places and cancels orders you would have placed by hand: a limit price, a size cap, a stop | Either | None | The rule was never sound; the bot repeats it faster |
| Providing liquidity | Rests orders on both sides and earns the spread when both fill | Maker | Optional | Only one side fills, just before news moves the price |
| Reference model | Computes a probability from outside data and trades when the market differs by more than costs | Usually taker | Weather forecasts, spot prices, other lines | The model is less accurate than the market |
| Cross-venue price gaps | Watches the same event on two venues and takes opposite sides when prices differ | Taker on both | A second venue's prices | The two contracts settle on different rules or times |
| Scheduled events | Reacts to a data release or game result within seconds | Taker | A fast feed of the result | Someone faster already moved the price |
Rule-based execution
This is the least ambitious kind and the one most people mean by a Kalshi trading bot strategy. The bot does not forecast anything. It enforces limits you set: enter only at or below a price, never hold more than a set number of contracts, stop for the day after a set loss. Hosted tools offer this category; see the best Kalshi trading bots compared.
Providing liquidity
A Kalshi market making bot quotes both sides of a market and tries to earn the spread. The fee schedule favors it, since resting orders are free by default. The risk is being filled on one side by a trader who knows more than the bot does, then holding that position into settlement, where a contract goes to $1 or to nothing. A liquidity bot needs inventory limits and a rule for pulling quotes before scheduled news.
Kalshi also runs a Liquidity Incentive Program that pays for resting orders in eligible markets whether or not they fill. Its help page limits it to most regular U.S. members and says the order book is sampled once per second and orders are scored on size and closeness to a reference price, with a daily reward pool of $1 to $1,000 per market, shared among participants. A sample counts only when both sides of the book hold the target size, the listed end date is January 1, 2027, and Kalshi can end or change the program at any time.
Reference models
Some markets settle on numbers that outside data predicts well: a city's daily high temperature, or whether Bitcoin finishes above a level. A bot can compute its own probability and compare it with the price. We publish that kind of calculation for weather and crypto markets, with its inputs (see how our reference probabilities are calculated), and score it against settled results on our accuracy page. Treat any model the same way: a gap between a model and the market means one of them is wrong, and it is often the model. The gap also has to be larger than the fee in the table above plus the spread.
Cross-venue price gaps
The same event sometimes trades at different prices on two venues. A bot can take opposite sides and wait for settlement. This is often called arbitrage, but the label hides the risks: the contracts may use different settlement sources or closing times, both venues charge fees, and one leg can fill while the other does not. Our cross-venue comparison lists matched markets and shows fees on both sides.
Scheduled events
Economic releases, game endings and weather observations arrive at known times. A bot can read the result and trade before the price adjusts. This is a speed contest against other programs, and Kalshi's rate limits cap how fast any account can send orders. Its documentation gives a new (Basic tier) account a write budget of 100 tokens per second, and most requests cost 10 tokens, so about 10 order requests a second. The highest tiers are earned from trading volume or assigned by Kalshi. For a solo developer this is the hardest of the five to do well.
What Kalshi's API terms allow
Kalshi publishes an API, and bots trading a member's own account are an expected use of it. The Developer Agreement sets the boundaries. To use the API you must be a member of the exchange and follow its rulebook. Section 3 says use of the API "is expressly limited to facilitating a members own trading on the Exchange; all other usages are disallowed and may result in account suspension."
The same section lists prohibited uses. Three matter for bot builders:
- Other people's accounts. "Facilitating trading or account creation by other members" is prohibited, and so is sublicensing the API to a third party.
- Data. Storing API data is allowed only for your own trading, and sharing it with third parties needs Kalshi's written authorization.
- Market integrity. Spoofing and wash sales are named as examples of prohibited use. A liquidity bot that places orders it never intends to fill is on the wrong side of this.
Other sections cover rate limits. Kalshi may set limits on API calls "at its sole discretion without notice", and may suspend access if it believes you are in breach.
The agreement does not say how the rule on other members' accounts applies to hosted services that run a bot through your own API key. If you plan to use one, ask Kalshi first. The tools in our guide to the best AI tools for Kalshi each handle this differently.
Where a Kalshi bot strategy can go wrong
- Automating an untested idea. Write the rule down, state what result would prove it wrong, and follow it by hand first. Kalshi offers a demo environment with mock funds for testing code, though its documentation warns that prices and market behavior there may not reflect real markets.
- Ignoring costs. A backtest that fills at the last traded price skips the spread and the fee. Use the bid and the offer.
- Thin books. A rule that works on 10 contracts may not fill on 1,000.
- Reading too much into a good month. A study of 588,000,000 Polymarket trades found the top 1% of users with positive profit captured 76.5% of profits, and a second paper found 3.14% of accounts statistically skilled (see sources). Both studied Polymarket, a different venue, so they show how concentrated results are in prediction markets generally and say nothing about Kalshi accounts in particular. Our article on who makes money on Polymarket has the detail.
- Operational failures. A crashed process with open orders, a leaked API key, or a rate-limit change can each cost more than a bad forecast.
Live prices for the markets most bots trade are on our crypto markets and climate markets pages.
Frequently asked questions
Are trading bots allowed on Kalshi?
Kalshi offers an API, and its Developer Agreement permits using it to facilitate your own trading as a member of the exchange. It prohibits facilitating trading by other members, sharing API data with third parties without written authorization, and manipulative orders such as spoofing and wash sales.
What is the best Kalshi bot strategy for a beginner?
Rule-based execution is the simplest to build and to check: the bot follows price and size limits you already use by hand. It does not forecast, so it only helps if the rule behind it holds up after fees.
Do Kalshi bots make money?
Some may, but a bot only carries out a rule. Fees, spreads and thin order books remove most small price differences, and research on Polymarket found profits concentrated in a small share of accounts. No strategy on this page comes with a promised return.
The short version: choose the strategy by its costs and its data, read the Developer Agreement before connecting anything, and test the rule before the program. For how Kalshi itself works, start with our Kalshi review.