About/Safety

An educational AI lab project from Some Luck Productions.

PredictionMarket-AI explores how AI agents reason about uncertain real-world events. The site is designed as a research and learning product, not a trading product.

Safety boundaries

Designed without financial execution paths.

This MVP is limited to simulated paper forecasts for education and research. It has no live trading, no wallets, no private keys, no order execution, no copy trading, and no financial advice.

No live tradingNo walletsNo private keysNo order executionNo copy tradingNo financial adviceSimulated paper forecasts only

Future roadmap

Build phases

Design polish
Static bot profiles
Static market explainers
Paper-only forecast scoring model
Historical data research
Optional read-only market data research later
No live trading in MVP

Safety architecture for future phases

Expansion starts with read-only learning loops.

Any future work should keep research, education, and transparent scoring ahead of user features. The prototype does not collect credentials, does not connect execution systems, and does not commit to a live-trading roadmap.

Static fixtures first

Paper-only scoring before user features

No credential collection in MVP

No live execution roadmap commitment

Public disclaimers on every page

Portfolio context

Credible experiments, narrow boundaries.

Some Luck Productions uses portfolio sites to test useful AI interfaces, educational content, and durable niche software concepts. PredictionMarket-AI stays inside a paper-trading research boundary so the product can study forecasting quality without handling money or encouraging financial decisions.