TL;DR
Forezai has released Polybot, an MIT-licensed open-source trading bot experiment for Polymarket. The project tests whether an AI agent can form probability estimates that differ from market prices, while warning that automated trading can lead to total capital loss.
Forezai has released Polybot, an MIT-licensed open-source experiment built for Polymarket that tests whether an AI agent can disagree with prediction-market odds and record why it reached a different probability estimate, a development aimed at AI forecasting research but framed by its creator as high-risk trading software.
The project was published at forezai.com/polybot.html and on GitHub as part of Thorsten Meyer AI’s Built in Public Day 13 of 19 series. According to the source material, Polybot opens the portfolio’s Markets family and is designed to compare a market price, such as a 62-cent contract implying about a 62% market probability, against an AI-generated estimate based on public information.
The stated workflow is narrow: estimate a probability, compare it with the market price, measure the gap, and decide whether the difference is large enough to clear costs and risk limits. The source material says the default action is no trade, with most markets skipped and any action limited to small, risk-capped positions when the disagreement is strongest.
Forezai describes Polybot as experimental software, not a financial product or profit system. The release says every estimate records its reasoning so a decision can be inspected after the fact, and it warns that prediction-market participation is restricted or prohibited in some jurisdictions, including for U.S. persons.
Polybot — when the AI disagrees with the odds
A prediction market puts a price on the future. Polybot asks: can an AI’s own estimate diverge from that price for real — and should it ever act on the gap?
Not financial, investment, legal or tax advice; not a recommendation or solicitation to trade, invest or use any software. Forezai · Polybot is experimental open-source software (MIT), provided “as is” without warranty of accuracy or profitability. Trading and automated trading carry a substantial risk of loss including total loss of capital; past or backtested performance does not indicate future results. Prediction-market participation is restricted or prohibited in some jurisdictions (including for US persons) — you are solely responsible for compliance with applicable law. Consult a licensed professional before any financial decision. Produced with AI assistance under human editorial oversight; independent commentary, the author’s own views. Product and company names are trademarks of their respective owners; mention does not imply endorsement.
AI Forecasts Meet Market Prices
Polybot matters because it puts an AI forecasting system against a live market whose prices already reflect the views and capital of traders. Prediction markets are hard to beat because prices aggregate information quickly; the release itself says skepticism should be the starting point for any system claiming to find an edge.
The project also shows a more cautious version of AI trading than many automated-bot pitches. Its published framing treats an apparent edge as a hypothesis rather than a fact, says backtests can flatter a system, and emphasizes that costs, market adaptation, and bad estimates can erase or reverse any apparent advantage.
For readers tracking AI agents, the release is less about a confirmed trading edge and more about auditability. The claim is that an agent can be used to study where its estimates diverge from market prices, with reasoning stored for review instead of hidden behind a black-box trade.
Use Claude to Build an AI Trading Bot: 90 Days with Stocks and Prediction Markets (AI Trading Bot Series Book 1)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Built In Public Day 13
Polybot was presented as Day 13 of 19 in Thorsten Meyer AI’s Built in Public series. The source material places it within an 18-product operator portfolio and identifies it as the first Markets node, alongside other product families in content, decisions, platform, defense, diagnostics, and readiness.
The project inherits several stated design principles from the wider series: local-first operation, provider-agnostic model use, and non-developer buildability. In practical terms, the release says the forecasting model is meant to be swappable, with no single model treated as an oracle about future events.
The release also stresses restraint. Its example table shows a trade only when an AI estimate differs enough from the market price to pass the system’s threshold, while smaller gaps or low-confidence estimates are skipped.
Use Claude to Build an AI Trading Bot: 90 Days with Stocks and Prediction Markets (AI Trading Bot Series Book 1)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unproven Profit Claims Absent
It is not clear from the source material whether Polybot has traded live capital, produced audited results, or generated any durable advantage over Polymarket prices. The figures shown in the release are described as illustrative of the logic, not a track record.
It is also unknown how the system performs across market types, how it handles stale or misleading public information, and how often model estimates would remain stable under different forecasting providers. The release does not provide independent performance testing, regulatory analysis, or user safety controls beyond the stated warnings.
open-source trading bot Polymarket
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Code Review And Testing
The next step for Polybot is public inspection of the MIT-licensed code and any future evidence on how its estimates compare with prediction-market outcomes. Readers considering the project should treat it as research software, check local legal restrictions, and consult a qualified professional before making financial decisions.
Further updates in the Built in Public series may show how Forezai connects Polybot to the rest of its Markets layer, but the current release leaves performance, adoption, and live-use details open.
How to Make Money Using AI in Options Trading: Proven AI Strategies, Algorithms, and Tools for Profitable Options Trading, Risk Management, and Automated Income Generation (AI For Investing)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Forezai Polybot?
Polybot is an MIT-licensed open-source experiment for Polymarket that compares AI-generated probability estimates with prediction-market prices.
Does Polybot prove AI can beat prediction markets?
No. The source material does not present audited trading results or a proven edge. It describes the project as experimental and says apparent gaps between AI estimates and prices are hypotheses.
Is this a recommendation to trade?
No. The release states that it is not financial advice and not a recommendation to trade, invest, or use the software. Automated trading can result in total loss of capital.
Can U.S. users participate in prediction markets through this tool?
The source material says prediction-market access is restricted or prohibited in some jurisdictions, including for U.S. persons. Users are responsible for checking applicable law.
What makes the project different from a standard trading bot?
The stated focus is auditability and research. Polybot records the reasoning behind each estimate and defaults to skipping markets unless a disagreement clears its stated threshold.
Source: Thorsten Meyer AI
