User repository on GitHub
Paper trading real Polymarket and Kalshi markets with a fake $100k: can Jev + Claude research beat the market?
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Demos & Experiments
Jev Paper Trader
An experiment that paper trades real Polymarket and Kalshi markets with a fake $100,000 bankroll, using TypeSafe's Jev as the judge of forecasts fed by Claude research; after about two weeks it reported being less accurate than the market's own price.
- Stars
- 3
- Language
- Python
- Reading time
- 2 min
- Published by
- Community
- Author
- MojoAI-King
- Last verified
- Oct 2026
Used inJev for Finance & Fintech
- Languages
- Python
- Capabilities
- Open sourceRequires API key
Highlights
- Paper trades real Polymarket and Kalshi markets with a fake $100,000 bankroll; it only reads public prices and places no orders.
- Jev judges forecasts fed with research by Claude, later ChatGPT and research agents, per the README's plan.
- Success was pre-registered in PLAN.md: profit or loss plus whether forecasts beat the market's own Brier score.
- Status 2026-10-09: after about two weeks every forecaster was less accurate than the market's price and strategies were losing.
- Run daily through VS Code tasks: check the Jev connection, preview live markets, then a daily run that settles and scans bets.
Watch out
README and GitHub metadata reviewed 2026-10-17; the excerpt truncates at the end of the daily-run instructions, so setup details, strategy code and the rebuild scope are not evidenced. No license is listed; last pushed 2026-10-17.
Questions about this entry
Short answers drawn from the linked project or page.
A paper-trading experiment on real Polymarket and Kalshi prediction markets with a fake $100,000 bankroll. It answers whether Jev, fed research by Claude, can grow the fake bankroll by more than luck would explain.
Jev acts as the judge of the forecasts: the README asks whether Jev, fed research by Claude and later other research agents, can beat the market. The project states it reads only public market prices and places no orders.
No: there are no accounts, no real money and no real orders. The $100,000 bankroll is fake and the project only reads public market prices from Polymarket and Kalshi to score its paper trades.
Two numbers decide it per the README: the profit or loss on the fake bankroll, and whether the forecasts beat the market's own price using the Brier score. Success and stop criteria were pre-registered in PLAN.md.
As of 2026-10-09, after about two weeks, every forecaster was less accurate than the market's own price and the strategies were losing. The measured diagnosis and the research behind a rebuild are in docs/REBUILD_SCOPE.md.
Open the folder in VS Code and use the tasks: first check the Jev connection, which should print Jev: OK, then preview live markets, then run the daily task once a day, which settles finished bets, scans for new ones and prints the report.
The repository was created 2026-09-27 and last pushed 2026-10-17 with 3 stars and no forks. No license is listed in the GitHub metadata in the digest, and the README documents no pricing.
More Jev resources like this
A demo that has three decision models read recent news and price Polymarket and Kalshi questions, then compares their probabilities with live market odds and replays how each model changed its mind article by article.
Jev Demos & Experiments#community#github#demo
Community
yorkeccak/model-vs-marketTypeScript
yorkeccak/model-vs-marketTypeScriptEleven small runnable projects for learning Jev and other System One decision models, each lesson pairing a one-command web demo with a whiteboard and README, from the three question types through jobs sorting and contract review to Jev inside a SQL query. Every demo makes live API calls.
Jev Demos & Experiments#community#github#python
Community
aianytime/jev-crash-coursePython
aianytime/jev-crash-coursePythonA small Python app that asks whether a headline is Onion-style satire or real news. Jev runs through TypeSafe's API with your own key and is compared side by side with AnyJev L0 and raw Qwen3-8B running locally, with an option to swap the answer order.
Jev Demos & Experiments#community#github#python
Community
andrisgauracs/onion-or-notPython
andrisgauracs/onion-or-notPythonTwo Jev demos over SprintPulse QA data: one scores release risk with a 10-level rubric evaluated against a reference score, the other diagnoses the main risk cause with a 9-label choice question, both sending the same per-ticket signals to Jev's System One API.
Jev Demos & Experiments#community#github#python
Community
avi350751/jev-qa-analyzerPython
avi350751/jev-qa-analyzerPythonStreamlit demo that turns live market data into structured investment decisions with TypeSafe Jev: yfinance supplies prices and fundamentals, then a 20-question model scores buy/sell conviction, financial health and risk.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
0xzee/jev-stock-decision-makerPython
0xzee/jev-stock-decision-makerPythonA runnable reference implementation of semantic data quality: it catches silent meaning drift that passes every deterministic BigQuery check, treats a semantic quality score as a first-class ops metric, and puts a circuit breaker at the gate.
Jev Demos & Experiments#community#github#python
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.