A market-facing proof that Core + Trust works.
Turn prediction-market probabilities into benchmarked, trust-aware signals.
Market Calibration Agent is a supporting asset and hero wedge for the Tollama stack. It combines market data, calibration scoring, and TSFM inference so teams can see when a market signal is reliable enough to inform research and risk workflows.
This wedge turns abstract market probabilities into benchmarkable trust signals. It is a concrete example of how Tollama Core evidence and Trust-style scoring meet in a real workflow.
Integrates REST APIs, WebSocket feeds, and subgraph queries from Polymarket to maintain a real-time, unified view of market data.
Computes Brier score, log-loss, and Expected Calibration Error per market, then turns them into composite trust signals with postmortem-ready audit trails.
YAML-defined alert rules trigger on probability deviation thresholds. Alert feeds are exposed as read-only API endpoints for downstream consumers.
Shows how Tollama Core-style TSFM inference can sit inside a guarded market workflow with rate-limiting, response caching, and circuit breakers.
Real-time monitoring dashboard showing market summaries, Trust Score distributions, active alerts, and calibration postmortems — designed for ops and quant teams.
Exposes market summaries, alert feeds, and performance postmortems as structured, read-only endpoints. No write operations, so it stays safe as a research and risk wedge.
Fully configurable via YAML — alert thresholds, model settings, API targets, and caching behavior are all declared without code changes.
Cleanly separated connectors, calibration pipelines, and API handlers. Add new market sources or scoring metrics without touching core logic.
Built with uv and a frozen lockfile for reproducible installs. The point is not another product line. It is a sharp, credible wedge for the Core + Trust story.
Use calibration metrics and TSFM signals to evaluate prediction market quality before incorporating them into models or strategies.
Monitor real-time alert feeds and Trust Score distributions to catch signal degradation before it affects downstream decisions.
Integrate the read-only endpoints into risk pipelines, dashboards, or agent workflows without turning the wedge into a separate platform.
Access structured postmortem data and calibration history to evaluate how prediction markets perform across different market regimes.
A supporting asset for teams that need to know when market probabilities are actionable.