TL;DR
Thorsten Meyer AI has introduced Forezai · Polybot, an MIT-licensed open-source experiment for comparing AI-generated probability estimates with Polymarket prices. The project is framed as a research tool, not financial advice or a trading recommendation, with automated trading risks and legal restrictions clearly stated.
Thorsten Meyer AI has introduced Forezai · Polybot, an MIT-licensed open-source experiment that lets an AI agent compare its own probability estimates with Polymarket prices and, in theory, act only when the gap clears strict risk limits.
The source material describes Polybot as a trading bot for Polymarket and as the first product in Forezai’s Markets family. It is presented as open source, available through forezai.com/polybot.html and GitHub, and built to test whether an AI system reading public information can form forecasts that differ from prediction-market prices.
The project’s central workflow is estimate, compare, decide. In the example provided, Polybot compares an AI estimate with a market-implied probability, checks whether the difference is large enough after costs and confidence limits, and defaults to no trade when the gap is too small or uncertain. The figures shown in the source material are described as illustrative, not as a performance record.
The announcement repeatedly warns that Polybot is not financial advice and is not a recommendation to trade, invest, or use the software. It also states that automated trading can lead to total loss of capital and that prediction-market access 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 Forecasting Meets Market Prices
Polybot matters because it places AI forecasting against a live market price, rather than a static benchmark. Prediction markets already combine the views and capital of traders, so any AI disagreement with that price is not treated in the source material as proof of an edge. It is treated as a hypothesis that can be inspected and can still be wrong.
That framing is central to the project’s news value. The announcement does not present Polybot as a money-making system. It presents it as an auditable experiment in whether AI agents can reason about probabilistic events, record why they disagree with market odds, and avoid action in most cases.
For readers following AI agents, open-source trading tools, or prediction markets, the launch is another example of AI systems being tested in domains where outcomes are measurable and errors can carry financial cost. The project also highlights the legal and practical limits around public access to prediction markets.
prediction market trading software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Built In Public Day 13
Polybot was introduced as Day 13 of Thorsten Meyer AI’s 19-part Built in Public series. The source material places it inside the Forezai operator portfolio and describes it as the first node in a Markets layer that also lists related concepts such as forecasting and trading.
The project inherits themes stated across the portfolio: local-first operation, provider-agnostic model use, and open inspection. In Polybot’s case, those ideas appear in the claim that the system can run on owned compute, can swap forecasting models, and records the reasoning behind each estimate.
The announcement also stresses a conservative operating premise: markets are hard to beat. The source material says a market price already aggregates information, opinions, and capital from traders, making skepticism the default response to any system that claims to find an edge.
AI forecasting tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
No Profit Record Provided
It is not clear from the source material whether Polybot has been tested with live capital, how it has performed in backtests, or whether any trades have been executed. The examples provided are explicitly described as illustrative of the logic and not a track record.
It is also unknown which AI models are being used, how market data is sourced, what exact thresholds trigger action, and how the system handles changing liquidity, fees, slippage, or adversarial market behavior. The announcement says the forecasting model is swappable and should not be treated as an oracle.
Legal availability also remains user-dependent. The source material says prediction-market participation may be restricted or prohibited depending on jurisdiction, including for U.S. persons, and says users are responsible for compliance with local law.
automated trading bots for prediction markets
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Repository Scrutiny Comes Next
The next test for Polybot will be public review of the code, assumptions, and safeguards now that the project is described as open source. Developers and observers can inspect how estimates are produced, how decisions are logged, and whether the no-trade default is reflected in implementation.
Future milestones will likely depend on whether the project publishes clearer documentation, reproducible tests, or performance data. Until then, the confirmed development is the release of an experimental MIT-licensed tool, not evidence that AI can reliably beat prediction-market prices.
open-source trading bot
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Forezai · Polybot?
Polybot is described by Thorsten Meyer AI as an open-source experiment and Polymarket trading bot that compares AI probability estimates with market-implied odds.
Is Polybot a recommendation to trade?
No. The source material states that Polybot is not financial advice and is not a recommendation to trade, invest, or use the software.
Does the announcement show Polybot making money?
No. The figures in the source material are described as illustrative examples of the decision logic, not a live or backtested performance record.
Why does the project focus on prediction markets?
Prediction markets attach prices to future events, giving AI forecasts a live benchmark. Polybot tests whether an AI estimate can differ from that benchmark in a way that is inspectable and risk-limited.
Can anyone legally use Polybot?
Not necessarily. The source material says prediction-market access is restricted or prohibited in some jurisdictions, including for U.S. persons, and users are responsible for local legal compliance.
Source: Thorsten Meyer AI
