How We Tag Wallets
Behavior becomes identity.
Wallets don’t come with bios — so we built a system that writes one for them.
Kairo’s AI watches how wallets move across time, chains, and categories. It learns their rhythms, risk profiles, and tendencies — then assigns a label that reflects who they are as a trader, not just what they’ve done.
These labels aren’t guesses. They’re generated by behavioral fingerprints pulled from thousands of datapoints per wallet.
🧬 Our Labeling Process
Here’s how it works:
Data Collection
Full trade history across chains
Token/NFT holdings (past & present)
PnL over time
Entry/exit timing per asset
Interaction types (swaps, LPs, mints, bridges, staking)
Behavioral Pattern Extraction
Holding durations
Position sizes and scaling
Rotation frequency between categories (e.g. memes → DeFi)
Reaction speed to market events
Volume preference and liquidity sensitivity
Fingerprint Formation
The model clusters wallets by behavioral similarity
Patterns emerge: some chase pumps, others build silent positions
These become trader personas
Identity Assignment
We assign a wallet up to 3 tags
Each tag reflects a dominant behavior cluster
Labels are updated dynamically — no one is locked into a single role forever
🏷️ What Tags Represent
A tag isn’t just a category — it’s a reputation signal.
Meme Swinger = rotates into low-cap hype tokens and exits early
Early Accumulator = consistently buys before major volume spikes
DeFi Farmer = high interaction with protocols, yield-focused
Cycle Exit Sniper = exits positions at local highs with precision
NFT Degen = active across multiple mints, flips often
These are just examples — we’ll walk through more in the next page.
🧠 Why Tags Matter
They help you instantly understand a wallet's personality
They let you filter by behavior, not just balance
They’re the foundation of real-time wallet insight via extension or API
They’re fully transparent: no label is hidden, no bias applied
Kairo doesn’t guess what a wallet is doing — it watches, learns, and labels.
Last updated
Was this helpful?