Retrieving the evidence ledger
Checking verified results
AI makes the prediction. SML freezes the evidence. Five Jev jury perspectives assess the published facts, with signed public results only after genuine TypeSafe model inference. Our own GME baseline miss stays visible.
Loading the current frozen-outcome ledger. Unavailable measurements are never filled with invented performance.
Reading the cryptographically signed jury verdict for the selected ticker. No simulated Jev votes will ever be displayed.
Five Jev perspectives: Causal, Historian, Skeptic, Microstructure, Tail Risk. Decisions assess research evidence only. TypeSafe AI is the third-party provider, not an affiliate or sponsor.
Checking verified results
Scores are descriptive, not independent trials or profit estimates. Each strategy faces the always-up baseline on its own exact future windows. Different tickers have separate scoreboards.
Submission access is initially restricted to the SML owner's server key. Self-service customer credentials and checkout are not represented as live until provisioned and verified.
Market-data researchers, agent developers, Discord communities, and analytics businesses can request a white-label Forecast Court integration. The core deliverable is the immutable claim → real-bar resolution → matched benchmark → public proof receipt pipeline, not a profitable signal guarantee.
Commercial package: scoped API access, branded judging and dashboards, research audit receipts, policy-gated baselines, and managed reporting. Pricing is quoted per integration; payment acceptance must be verified separately.
Currently available for the mature GME 36-outcome cohort. New tickers must accumulate at least 20 actual resolved five-day predictions before their audit is eligible.
Base mainnet USDC, x402 v2. Clicking presents machine-readable payment requirements; it does not authorize a wallet transfer. Settlement and accepted delivery require a compatible paying agent. Public Court evidence is always free.