Best Prediction Market Bots (2026): Kalshi & Polymarket

A prediction market bot, or prediction market trading bot, is software that places or manages orders on an event-contract exchange such as Kalshi or Polymarket, following rules instead of a person clicking. Before you compare bots, answer three questions. Which exchange can you actually use? Will you write code? Does the tool place orders, or only show data? This guide sorts the real options by those three answers.

Who wrote this. We build and sell Bot for Kalshi, a hosted bot builder that trades Kalshi only. It does not trade Polymarket or any other exchange. This is a publisher-disclosed buyer's guide, not an independent ranking, and one of the options below is ours. We list it first in each section because it is the one we can describe from the inside; the order is not a ranking.

Start with the exchange, not the bot

A bot can only trade where you hold an eligible account with programmatic access. That one fact rules out most of the market for most readers, so settle it first.

Kalshi and Polymarket can list markets on the same event and still settle them differently, because each writes its own rules, sources, and edge cases. A bot built for one exchange rarely moves to the other unchanged, and a "multi-venue" label tells you nothing about how well each side works. Our Kalshi vs Polymarket comparison covers the structural differences.

The five jobs a prediction market bot does

"Prediction market bot" covers five different jobs, and each one fails in its own way. Name the job before you pick the tool.

Go deeper on each job: AI in prediction markets, what is real in Kalshi arbitrage, market making, and copy trading on Kalshi, where the public trade feed is anonymous and Kalshi's social profiles and leaderboard do not copy orders into your account.

Hosted bots that place orders for you

These run on the vendor's servers, so you don't host any code. You give them trading access to your account and write the strategy in a builder or in plain English.

Bot for Kalshi (our product): Kalshi only

You describe a rule in plain English or build it step by step. The rule stays visible and editable, and the hosted engine runs it against your own Kalshi account, with limit orders by default. You can run it in Paper mode against live prices before using real money, and set account loss caps, position limits, and stop conditions. It does not trade Polymarket or offer historical backtesting, and paper results do not reproduce live fills. Bot for Kalshi has one Complete plan at $99 per month; payment is required and there is no free trial. Our Kalshi bot options guide compares it with other Kalshi tools in more depth.

Turbine: Kalshi and Polymarket

Turbine (TurbineFi) turns a plain-English strategy into a bot, backtests it, and deploys it to a dedicated runner. Its product pages name Kalshi and Polymarket, and its docs describe live and paper modes. As checked October 22, 2026, Turbine's pricing page lists three plans at $9, $99, and $199 a month, with 10, 100, and 400 deploys per week and one, three, or five bots at a time. Its runtime page says runner compute on Starter and Basic is billed separately, at cost. Backtesting and paper trading are listed from the entry plan up. A backtest is a historical simulation, not a live result. Our Bot for Kalshi vs Turbine comparison covers the differences.

Elastics: Kalshi and Polymarket, free, in private beta

Elastics describes itself as an AI trading assistant that connects to Kalshi with API keys and to Polymarket with a wallet, with agents you can simulate before they go live. Its FAQ says Elastics is free and non-custodial. Its app says it is in private beta, with access granted in batches. We have not tested its execution.

Public: Kalshi contracts inside a brokerage account

Public, the brokerage app, offers event contracts through Kalshi. Its prediction-markets page says its AI Agents can monitor prediction markets and act on rules you write in plain English, and that you review and approve the plan before an Agent runs. This is a brokerage route, not a bot you connect to your own Kalshi account. The same page says event contract accounts are not SIPC protected, and Public's general Agents page notes that some capabilities are not yet available.

Polymarket copy-trading bots

Many of the hosted "Polymarket bots" we found are copy-trading or sniping bots run through Telegram or a web app, such as PolyCop, PolyGun, and Polycopy. Their own pages describe per-trade fees of 0.5% to 1%, and Polycopy's automatic copy bots also require a $30-a-month paid subscription. They serve Polymarket's international exchange, which its documentation lists as close-only for the United States. Because Polymarket positions sit in a crypto wallet, find out exactly where the private key lives and who can sign with it before you deposit anything. We have not tested any of them.

Bots for one Kalshi market family

Some products automate a single family of Kalshi markets rather than any rule you write. They can be simpler to start with, but check that they cover the market you actually want to trade.

• Reflex trades Kalshi's crypto 15-minute and hourly markets from a browser extension and mobile web app. As checked October 20, it has a free tier and a paid tier at $14.99 a month, with paper trading on the paid tier.
• EventEdge automates Kalshi sports markets using its own probability models. Instead of a subscription, it charges a 1% volume-based fee drawn from prepaid credits.

Short-interval crypto markets are a common target for bot products, and fees decide them. Across 6,298 settled KXBTC15M windows (Kalshi's 15-minute Bitcoin series), we measured a gross edge of about +0.67 cents per contract for buying the leading side, against an average taker fee of about 1.55 cents, so the taker fee is larger than the measured edge. Our guide to how Kalshi's 15-minute Bitcoin markets work covers the market itself, and the Bitcoin bot walkthrough shows the full example.

Code you run yourself

If you write code, you get full control and no subscription. You also own the hosting, retries, monitoring, key security, and the work of reconciling fills against the exchange.

Official Kalshi API and SDKs

Kalshi publishes official SDKs for Python and TypeScript, and a demo environment with separate credentials so you can test order logic before production. Our Kalshi API guide covers signing, rate limits, and WebSockets, and the Python bot tutorial walks through a working bot.

Polymarket's official SDKs

Polymarket's developer docs list two current SDKs: @polymarket/client for TypeScript and polymarket-client for Python. The older py-clob-client repository was archived in May 2026, and its README says the client no longer works, so skip tutorials built on it. Polymarket's own open-source AI agents framework was also archived in May 2026; its code last changed in November 2024. Polymarket US has its own API and SDK, separate from the international exchange.

pmxt: one API for many exchanges

pmxt is an open-source (MIT) library that puts Kalshi, Polymarket, and other prediction markets behind one Python or TypeScript interface. Its README says hosted order placement covers Polymarket, Opinion, and Limitless today, while Kalshi orders go through the self-hosted library with your own keys. Its hosted data API has a free tier and paid tiers at $29.99 and $99.99 a month. A shared interface saves plumbing, but you still have to check that two markets with the same name settle the same way.

SimpleFunctions: command line and SDK for agents

SimpleFunctions offers a command line, SDK, and an MCP adapter for AI assistants, for prediction-market research and execution. Its docs say its order commands can place or cancel real-money orders on Kalshi or Polymarket, and that they refuse to run unattended unless you explicitly allow it. Its pricing page lists a free tier and paid tiers at $20 and $200 a month.

Open-source AI bots for Kalshi

Two widely shared MIT-licensed projects are ryanfrigo/kalshi-ai-trading-bot and OctagonAI's Kalshi trading CLI. Both had commits in early October 2026. The first warns that its example strategies lose money on some markets and calls itself a toolkit to fork, not a turnkey bot. That kind of candor is a good sign. For the wider field, our bot catalog compares open-source Kalshi bots side by side; stars measure attention, not quality.

Research tools that are not bots

Many tools in "best bot" lists only show data. They can be useful upstream of a bot, but they don't place orders.

• Oddpool compares Kalshi and Polymarket prices. Oddpool's pricing page says it is no longer accepting new subscriptions; its free tier still lists dashboards, cross-venue search, and paper trading.
• Dome, a unified prediction-market data API, says in its docs that it was acquired by Polymarket and that its APIs were scheduled to reach end of life on April 28, 2026. Don't start a new build on it.
• Signal channels and dashboards that post picks are research at best. Our guide to prediction market signals covers what to check before acting on one.

Comparison table

How to vet any prediction market bot

1. Does it place orders, and where? Ask which exchange, which markets, and whether you are eligible to use that exchange. Dashboards and signal channels are not bots.
2. What access does it hold? A Kalshi bot normally uses an API key you create and can delete in your Kalshi account. Bots on Polymarket's international exchange sign with a crypto wallet; Polymarket US uses API keys instead. Never paste a wallet private key or seed phrase into a website or chat bot: whoever holds it controls the funds.
3. Can you cap the damage? Look for limit orders, a daily loss cap, position limits, and a pause control. In some products a pause stops new actions but does not close open positions, so check which yours does.
4. Can you test it first? Kalshi's demo environment, a product's paper mode, or a dry run should come before real money. Paper results still won't match live fills exactly.
5. Does it log every decision? Each order should show the time, price, quantity, and the rule that fired. You need that to debug, and for taxes.
6. Does it account for fees? Many short-term rules look profitable before fees and lose after them. Check an order in the prediction market fee calculator.
7. How does it talk about results? Backtests, leaderboards, and screenshots are not live track records. Treat a promised return or a fixed win rate as a red flag.
8. Who maintains the code? For open source, check the license, the last commit, and the open issues, and read the order path before you give it a key.

Our bot risk-management guide covers these controls in more detail.

What no bot can do

A bot enforces a strategy faithfully. It does not create an edge the strategy lacks. On liquid, popular markets you may be trading against other automated traders, and fees are charged whether the rule is right or wrong. Our look at how many bots trade on Kalshi explains why that share can't be measured from the outside. Traders can lose money with or without automation, including their full stake on a contract.

Which one fits you

• You trade Kalshi and don't code: a hosted Kalshi builder, run in paper mode first. Compare ours with Turbine, and check whether your market is better served by a single-market bot.
• You trade Polymarket: we can't help you execute there. Start with Polymarket's official SDKs, or a multi-venue builder if you don't code.
• You want one codebase for both: a unified library such as pmxt, with your own settlement checks and risk limits on top.
• You want AI: let the model draft or classify, and keep sizing, limits, and execution deterministic. Our guide to LLM signals shows the pattern.
• You only want to compare prices: a research tool, not a bot.