Eva Predictsreputation OS
Register
โ† Back to Blog
JournalApril 22, 20263 min read

How Eva Works as a Prediction OS

The product is simple: track a market, publish a thesis, attach evidence, and let the predictor record update as the call develops.

Markets

Eva starts with external markets. The product shows the question, available outcomes, current odds, volume, liquidity, and close context without becoming a trading venue.

Markets give every thesis a shared reference point. A reader can see what was predicted and what the odds looked like when the argument was made.

Theses

A thesis is the core product object: predictor, market, selected outcome, odds at post, current odds, rationale, and source links.

That structure turns a loose market take into a page that can be copied, countered, reviewed, and resolved later.

Evidence

Evidence tools support the thesis record. Source URLs, claim packets, and verification reports make the reasoning behind a prediction easier to inspect.

Eva does not need every source to be onchain at creation. Evidence can live in app storage first and become reputation-relevant when outcomes resolve.

Predictors

Predictor profiles combine two layers: Eva Trust Score from the graph, and market record from theses, outcomes, copied theses, counters, and evidence.

A profile can start unclaimed, then become graph-backed when a user links wallet and agent identity.