Metaforecast

Metaforecast is an experimental project developed by the Quantified Uncertainty Research Institute (QURI), a non-profit organization dedicated to improving the quality of public forecasting and world modeling. It functions as a specialized search engine that aggregates data from over 10 different forecasting platforms and prediction markets, including Metaculus, Polymarket, PredictIt, Smarkets, and Good Judgment Open. The tool was created to solve the "fragmentation problem" in the forecasting ecosystem, where valuable probabilistic data is often scattered across various websites with different interfaces and formats.

By fetching data daily, Metaforecast provides a unified view of thousands of active forecasting questions. One of its most distinctive features is a subjective "star rating" system (1–5 stars) used to estimate the reliability or "robustness" of a specific forecast based on factors like the number of forecasters, trade volume, or market liquidity. As an open-source tool, its codebase—including the scrapers used to fetch data and the frontend website—is publicly available on GitHub, allowing the community to audit the aggregation logic or integrate the data into other services via a JSON endpoint.

From a technical standpoint, it operates as a data aggregation and normalization layer, pulling structured forecasting data and presenting it in a unified interface. Being open source, it allows developers to extend integrations, improve models, or build custom analytics pipelines on top of its dataset.

It is particularly valuable in domains where uncertainty and probabilistic reasoning are critical, such as economics, geopolitics, technology trends, and emerging risks.