Loading
Loading table data
Documentation
Launch DNA reads the opening minutes of a Pons V2 launch on Robinhood Chain and turns them into something you can scan. This page explains what it measures, what it deliberately does not claim, and where the data currently comes from.
In the first minutes of a launch there is almost no price history to read. What there is instead is structure: how quickly the first buyers arrived, who those buyers were, how fast the bonding curve filled, how concentrated the supply became, and how closely the whole shape resembles launches that have already played out.
Launch DNA collapses that structure into one number out of 100, then shows the five components behind it so the number can be argued with rather than taken on faith. Every panel on a token page is there to let you disagree with the score.
The terminal is descriptive, not predictive. It reports what a launch looks like and what similar launches did afterwards. It never tells you what will happen next, and it never tells you what to do.
The headline DNA score is a composite out of 100, split into five components. Each is also reported on its own, because a single high number can hide a weak component — a launch can fill its curve at record speed and still be held by four wallets.
How quickly the first buyers arrived and how fast the bonding curve filled in the opening minutes.
The share of early buyers classified as historically strong wallets by wallet intelligence.
How much of the early flow looks independent rather than bundled or repeated from the same cluster.
How concentrated the supply is across the top holders relative to the rest of the buyers.
How closely the opening shape of this launch resembles launches already in the archive.
These bands exist so a number has a word next to it. A band is a label, not a grade, and not a verdict on whether a token is worth buying.
The opening minutes of a launch are described as a set of features — entry cadence, buyer mix, curve fill rate, sell pressure — and compared against launches already in the archive. The closest matches are listed with what actually happened to them.
That framing is deliberate. “Launches that opened like this went on to do X” is a statement about the past. Cohort statistics — graduation rate, median peak, median drawdown, median time to peak — are reported next to every match list so the spread is visible, not just the best case.
A high similarity score means a launch resembles its cohort. It does not mean it will follow the cohort.
An address that repeatedly appears early in launches that go on to graduate behaves differently from one that buys in the launch block and sells into the first wave of organic buyers. Both patterns are visible in transaction data without knowing anything about who owns the address.
Every early buyer carries a marker, and every marker is paired with its meaning in writing. A coloured dot on its own communicates nothing to a reader who cannot see the colour, or who has never been told what it stands for.
A new wallet is explicitly neutral. No history is not the same as bad history, and the terminal does not imply otherwise.
This build serves the Launch DNA preview dataset: a fixed, self-consistent set of launches, wallets, transactions and archived outcomes used to exercise the full interface while the live indexer is being connected. Figures are stable across reloads by design, so a view you share looks the same to whoever opens it.
Ages and timestamps are measured against a fixed snapshot of Sep 10, 2026, 14:32 UTC rather than the wall clock, which is why the terminal renders identically on the server and in the browser.
Until the live feed is switched on, treat every number here as an illustration of the product rather than a live measurement of any specific token.
The interface never reads a dataset directly. Every page and component goes through a single typed module, lib/api.ts, which returns an ApiResult<T> envelope. Switching from the preview dataset to the live indexer is a change in one file, and no screen has to be rewritten for it.
If the feed is switched on before the endpoint exists, the terminal does not attempt a request: every screen renders a typed “data source unavailable” state instead, so a misconfigured deployment fails visibly rather than silently.
Questions, corrections or a launch you think the index is missing — reach us on X or Telegram.
Back to the terminal overview or the live launches feed.