Reading a verdict
Every news group gets a verdict. It is recalculated each time a new token joins the group, so it can change in the first minutes.
What we analyze. Only groups born from a post on X: the tokens link to the post, or we found the post they came from. A group with just a shared ticker or name and no post behind it gets no verdict and is not shown. Tokens launched in pump.fun Mayhem Mode are left out completely: they never join a group, are never OG or vamp, and are not counted in stats or accuracy.
Group status
| Status | What it means |
|---|---|
| Confirmed (OG) | The engine settled on one token. This is the real one. |
| Early pick (early leader) | Few tokens or little data so far. The leader can still change. |
| Too close (disputed) | Two or more tokens score almost the same. All of them are named. |
| No real token (no OG) | Nothing in the group looks like the real one yet. Every coin so far looks like a copy. |
Token roles
| Label | What it means |
|---|---|
| Real (OG) | The token the verdict picked. |
| Copy (vamp) | Another token under the same news after the OG was confirmed. |
| Contender | One of the close tokens in a "too close" group. |
| Pending | Not decided yet: the group is still forming. |
| Fake letters | The ticker uses look-alike characters to imitate another one. It is never the OG. |
| Dev sold early | In the first minutes the creator already sold for at least twice the SOL they put in. Such tokens usually die in one red candle, so we never call them the OG. |
| Banned dev | The creator wallet was banned by Unvamp moderation. Its tokens are never the OG. |
Score and reasons
The OG has a score from 0 to 100: how sure the engine is. Next to it are two to four short reasons in plain words, for example "market picked it", "ticker matches the post", "first with the post link".
How the OG is picked, in short: Unvamp reads the post and the tokens (name, ticker, picture, links, timing) and how the market treats each of them, and an AI model and a set of rules score every token. We do not publish the exact criteria, so that copy launchers cannot tune their tokens to them.
Accuracy
The dashboard has a public Accuracy page. It compares the pick made a few minutes after the first token with the token the market ended up choosing, and with the naive strategy of buying the first token launched. The numbers there are live, not curated.
A verdict is automated analysis, not financial advice. The OG is the token that matches the news, not a promise that it will go up.