EVIDENCE, NOT INTENT

Method

Unbundle reconstructs the first 150 slots and explains every classification with links to the underlying transactions.

01

Shared blockhash

A non-nonce transaction reuses the create transaction's recent blockhash.

02

Shared lookup table

A buy uses a non-public address lookup table also used by the create transaction.

03

Co-signers

Wallets sign the same transaction together.

04

Same-slot batch

Transactions share a slot, blockhash and priority fee with a dev-linked transaction.

05

Temporary holder

Tokens move from a dev-linked wallet to another wallet inside the launch window.

06

Front → back handoff

A linked wallet exits most of its position while at least two linked wallets buy in the same slot.

07

Coordinated group

At least three wallets share strict transaction fingerprints without a verified dev link.

08

Outside sniper

An early buyer uses a durable nonce or has no shared dev/group fingerprint.

09

Migration

PumpSwap migration instructions and pool creation are excluded from holder layers.

Why naive block-0 counts mislead

Independent sniper bots are normal at launch. Counting every first-block buyer as part of a creator bundle produces a larger number without proving a relationship. Unbundle separates those wallets using blockhashes, lookup tables, signers, fees, nonces, tips and instruction layouts.

Tools built to hide links—for example exchange-funded wallets with time gaps—can avoid some checks. Results are evidence of coordination, not proof of intent.