VIRGIL is an autonomous agent that grades every launch on the Virtuals Unicorn launchpad — reading each one's on-chain configuration against a single calibrated, public rubric. It runs on its own, around the clock. No judgment calls, no black box: the same input always yields the same reading, and anyone can check the instrument.
Most launches are not projects — they are inventory. A single deployer mints tokens by the thousand, minutes apart, around the clock. VIRGIL grades each by what was configured on-chain, not by what it was named.
Every launch returns a reading: a composite score and the three measured signals — configuration, creator history, declared presence — each with the plain reason for its value.
VIRGIL does not forecast price. A high reading means a founder enabled every honesty signal available — not that the token will rise. The instrument reports the state at launch. Nothing more.
Sniper-tax decay, capital formation, and the 60-day founder commitment. On-chain choices that cost a serious founder nothing and a fleeting one everything.
The wallet's age, transaction record, and prior launches. A fresh address on its four-hundredth token reads nothing like a real builder's.
A verified account, a substantive description, real documentation. Anonymity isn't disqualifying — the absence of any claimed identity is.
No founder can pay to raise a reading, suppress one, or be omitted. The paid product is depth of analysis — never the grade.
A wrong reading is corrected as a correction, logged and visible. The instrument is never silently recalibrated.
The rubric is published and versioned. Apply the same weights to the same launch and the reading reproduces exactly.
Because the rubric is deterministic and public, an agent can verify the method once and rely on the stream. The grade is a calibrated, checkable reference — not a black box to be trusted on faith.
Read the methodology →