Calibrated foresight, grounded in capital
A continuously updated causal model of world state. Ask any question and get a streamed probability with the drivers and markets behind it.
Ask any question about the future
Get back an answer you can act on, with the markets and drivers behind it.
Forecast accuracy improvement
Walk-forward backtests across 17 assets. Median improvement over industry-standard volatility forecasting models across equity, rates, FX, and commodities.
Explicit causal reasoning
We fuse market data with news, and your own private sources, then apply an explicit causal model to answer questions no single market prices, including conditional and counterfactual ones.
Auditable probabilities
Streamed probabilities with a full audit trail, traceable to the contributing sources and weighted by confidence.
Explore Situation RoomsWorld events are connected
COVID → Supply Chain Issues
Russia Invades Ukraine → Energy Shock
Fed Pivot → Credit Tightens
SVB → Credit Suisse → Credit Crunch
Trump Returns → Tariff Risk
DeepSeek → Tech Sell-off
Strait of Hormuz → Crude Spike
PolyBridge Features
Ask beyond any single market
Ask any question about the future. PolyBridge searches across markets and venues to answer questions no individual market covers.
Probability with uncertainty
Get a calibrated probability with an uncertainty range, not just a single number.
Traceable reasoning
See the events and signals driving the result, follow the evidence trail, and inspect the exact source markets and venues used.
Scenarios, live updates and alerts
Change assumptions to see how probabilities move, track updates as markets change, and set thresholds for notifications.
Adjust an input. See the downstream effects
“If the Fed holds in September, what happens to emerging-market sell-off probability?”
The engine works across 170,000+ markets and returns a quantitative answer with every contributing market visible.
The interface
A probability you can act on
Every answer is a precise number with a credible interval. The interval widens when market evidence is thin, or stale, and narrows when it is strong.
$151.8m in real capital backing this answer
Each probability traces to the markets that informed it, weighted by liquidity, where real capital is at risk.
31% Credit Stress. 19% Geopolitical Escalation
Each driver is a node in the causal model. Condition on it, intervene, or test a counterfactual.
Common questions
How PolyBridge fits into forecasting, monitoring, and decision work.
PolyBridge returns streamed probability distributions. Uncertainty is explicit. Causal drivers behind the result are reported. Quantitative answers are traced back to source data, from prediction markets and other sources.
Yes. PolyBridge answers have mechanistic explainability by design, rather than narrative explainability.
The volume and number of event contracts on prediction markets is growing, but the number of questions will always be finite, unlike the number of questions worth answering. This is a gap PolyBridge fills.
LLMs excel at providing compelling narrative, whereas PolyBridge returns streamed quantitative probabilities conditioned on markets. The use cases are different and complementary.
PolyBridge maintains a continuously updated latent model of world state. It's mechanistically interpretable and supports do-calculus on named concepts.
We are continuously adding new data sources, which improves model calibration with every resolved event.
Markets aggregate the views of many participants, including other models, into a price. The best models who have an edge are incentivised to participate and publish their otherwise private information. That makes them the most efficient single place to start.
No, this is the problem PolyBridge causal world model solves. Generalisation to low and no liquidity regimes.
Every answer is a structured probabilistic output. It contains:
- A calibrated probability with a confidence interval. A range that reflects uncertainty, sized to the strength and freshness of the underlying evidence.
- An evidence trail. Each connection is legible. You can see which upstream events drive the answer and by how much.
- Source markets with contribution weights. Every market that informed the answer is named, with its influence shown explicitly.
- A public-data prior, separate from market signal. What historical reference data implies versus what live markets are pricing. Divergence between the two is shown.
- Support and freshness metadata. How many resolved historical examples back the estimate, and when the underlying data was last updated.
This structure means every answer can be reviewed and defended.
Additional documentation is available to qualified institutions upon request. Contact PolyBridge
Teams making high-stakes decisions under uncertainty across defense and intelligence, long-horizon investing, financial services, energy and commodities, and corporate strategy. They use Situation Rooms to monitor live crises, the Forecast Interface to stress-test decisions, and the API to embed foresight into their own systems.
Request access for Situation Rooms, the Forecast Interface, Search, Forecast, and approved RVOL data endpoints, agent workflows, or other PolyBridge tools.