The short version
Eva Protocol is a prediction and verification reputation layer on Avalanche. It helps people and agents turn a public market call into a structured record: market, outcome, odds context, rationale, source claims, evidence, and eventual outcome history.
The core idea is simple. A post can disappear into a feed, but a thesis page can be revisited. A claim can be reused across markets. A curator or predictor can build a track record that outlives one thread.
The product object
The main object in Eva is the thesis. A thesis points at an external market, names an outcome, captures the reasoning behind the call, and links evidence that readers can inspect.
Theses can be copied, countered, shared, and resolved. That gives prediction discourse a product surface without requiring Eva to become a trading venue.
The trust layer
EvaTrustGraph is the long-lived trust primitive. It is where identity, curator registration, self-stake, and graph-backed reputation belong.
The app can start with offchain thesis and claim records, then promote durable resolved outcomes into reputation when the adapter boundary is ready.
What is live today
Eva has a live app surface for markets, theses, predictors, claims, verification, blog content, and curator onboarding. The canonical production target is Vercel, and the primary chain is Avalanche C-Chain.
Eva does not execute trades in v1. It does not custody funds or run a native prediction market today. It links to external venues and keeps the reasoning record useful.