Who is really trading this token?
Feed it a mint address. BotMeter pulls the token's newest ~300 swaps and runs them through ten detectors: twin trades in one block, buy-then-sell in a single slot, sub-second bursts, cloned order sizes, MEV tags and more. Out comes the machine share of both trade count and dollar volume, plus every raw tally behind it.
Awaiting target. Paste a mint above or tap one of this hour's busiest tokens.
How the score is built
One odd trade proves nothing on its own, so detectors are weighted and combined. Every count is printed so you can check the reasoning yourself.
01Input sample
Roughly the 300 newest swaps from Jupiter's open trade feed: wallet, direction, size and block. On a hot token that is a few minutes of activity; on a sleepy one it can stretch over hours, and the exact window is shown with the result. Up to 20 of those transactions are also fetched in full from public Solana RPC nodes.
02Strong detectors
Jupiter tagging the swap as MEV; one wallet buying and selling inside the same block (an atomic loop or sandwich leg); one wallet landing two separate transactions in a single ~0.4 s slot. Any single hit classes the trade as automated.
03Medium detectors
A swap within 3 blocks (~1.2 s) of the same wallet's last one; a wallet with 12 or more swaps in the sample; one exact SOL amount used 3+ times by one wallet. Two of these, or one plus two weak hits, means automated. A lone medium hit is labelled "possible".
04Weak detectors
Tidy sizes like 0.1, 0.5 or 1 SOL (trading apps use these as presets too), a Jito tip, a priority fee at or above 1M micro-lamports per compute unit, and a top-level program missing from our known list, which might be a private bot contract or just an app we haven't catalogued. These only matter in combination.