What changed
The original Eva MVP put article verification at the front of the product. That made sense as infrastructure: fetch a source, extract factual claims, score the evidence, and connect the result to a trust graph.
The new MVP changes the front door. Eva now starts with prediction theses: a market, an outcome, the odds at the time of posting, the rationale, and the evidence behind the call. Verification remains useful, but it becomes support for a thesis instead of the main loop.
Why X first
The target audience is already on X. They post screenshots from prediction markets, explain why odds are mispriced, argue in replies, and build informal reputations through threads that disappear into the feed.
That is the GTM opening. Eva should not ask that audience to adopt a new behavior before they get value. It should turn the behavior they already have into a public track record they can share, defend, and compound.
What stays
EvaTrustGraph remains the important primitive. A predictor can start as an unclaimed X profile, but the stronger version is graph-backed: wallet, ERC-8004 agent identity, stake, and reputation that can outlive a single thread.
The protocol still cares about durable trust. The product just gets there through a sharper wedge: visible market calls with outcomes people can inspect.
What v1 does not do
Eva does not execute trades in v1. It does not custody funds, place orders, or run a native prediction market. Copy actions are previews and external links only.
That boundary matters. The MVP is testing whether predictors want public reputation for their calls before the protocol takes on settlement, execution, or deeper market infrastructure.
What the MVP tests
The metric is not page views. The metric is weekly active predictors: people who publish calls, come back, and care enough about their record to keep using Eva.
If that loop works, the next layer is obvious: resolved outcomes can feed reputation, counter-theses can improve signal, and the trust graph can become the durable memory behind prediction-market discourse.