Kalshi AI trading agents just moved from developer side projects into a mainstream brokerage app. On September 24, 2026, investing platform Public announced a partnership with Kalshi that lets everyday users set up automated strategies tied to prediction market odds. Instead of manually checking whether the Fed will cut rates or whether a company beats earnings, users can now hand that job to an AI agent.
Public calls itself “the world’s first agentic brokerage.” The company is backed by Accel and Tiger Global, and it already manages billions of dollars for retail investors. Kalshi supplies the event contracts and handles execution. Public supplies the interface and the automation layer on top.
How the AI Agents Actually Work
The setup is simple on the surface. A user tells Public’s AI what they want in plain language. The agent then converts that request into fixed, rule-based logic — not an open-ended AI that trades however it wants.
Public’s co-CEO Leif Abraham described it as automation for investors who already build their own portfolios. According to the company’s official launch announcement, every rule is shown to the user as a visual plan before it goes live. Nothing executes until the person approves it. Once approved, the agent can only act within the exact conditions that were set — it cannot improvise new trades.
Three types of actions are on offer. Agents can send alerts when odds move past a chosen level. They can trade Kalshi contracts directly. And they can trigger actions elsewhere in a user’s portfolio, like buying a stock if a related prediction market probability crosses a threshold.
Real Examples From the Launch
Public gave a few concrete use cases in its release. An agent can buy shares in a healthcare company automatically if the market-implied odds of FDA approval rise above a set number. Another can watch rate-cut probabilities and send a summary of a user’s bank stock exposure if those odds shift sharply. A third can buy put options if the implied probability of an earnings miss climbs past 60%.
These examples matter because they show the product isn’t limited to trading Kalshi contracts for their own sake. It uses Kalshi’s odds as a data feed that can move money in a completely different asset — stocks, options, whatever the user already holds.
Why Kalshi, Specifically
Kalshi is the largest prediction market operating under direct CFTC oversight, which is exactly why a mainstream brokerage picked it as a partner instead of building this on an offshore or unregulated platform. Public’s release describes Kalshi as “a CFTC-regulated financial exchange.” That regulatory status is what lets a regulated brokerage plug into it without extra compliance headaches.
The market categories covered are broad: crypto, commodities, climate, economics, corporate events, macro indices, tech and science, and politics. That range means the AI agents aren’t just a political-betting gimmick — they’re pitched as a general tool for translating any Kalshi-tracked event into a portfolio action.
What This Means for Retail Traders
Automated trading tied to prediction markets is not new. Developers have been building bots against Kalshi’s API for a while, and a search for the topic turns up dozens of GitHub projects and technical blog posts from people running their own scripts. What changes with Public’s launch is distribution. A retail investor no longer needs to write code or manage API keys — they describe a strategy in plain language, review the resulting rule, and turn it on.
That accessibility cuts both ways. Handing a strategy to an agent removes friction, but it also removes the moment-by-moment judgment a human trader would normally apply when a market gets volatile. Public’s approval step is the main safeguard here: nothing fires without the user signing off on the rule first, and the rule can’t be changed mid-flight without going back through that review.
For readers who want to build similar automated setups without going through a broker’s packaged tool, our guide to the Polymarket API covers the basics of connecting programmatically to prediction market data, which applies conceptually even though it’s a different platform. And if the appeal here is following what already works rather than building new rules from scratch, our piece on copy trading on Polymarket explains a simpler alternative: mirroring an established trader’s positions instead of scripting your own conditions.
The Bigger Trend
This launch fits a pattern that has been building through 2026: prediction markets moving from a niche trading venue into infrastructure that other financial products plug into. Kalshi has spent the year signing partnerships that push its contracts into places retail investors already are, rather than asking traders to come to Kalshi directly. An AI agent that trades a Kalshi contract because it read a Fed odds shift is a step further down that road than a simple data feed or embedded widget.
Whether this becomes a mainstream habit depends on something simple: whether the agents make money for the people using them, or just make it easier to lose it faster. Public’s rule-based, approval-gated design suggests the company is aware of that risk. The bigger test comes once volatile weeks — a surprise Fed decision, a contested election night — put the automation under real pressure instead of a calm market.
