Cluster Maps (Coming Soon)

Outside of our Cognitive Framework and AI-Auditing tool, we embedded all of the features that make this an end-to-end platform for token safety analytics. We built our own clustering software that can detect clusters of token holder wallets that have been obfuscated by second or third-order transactions. Our clustering algorithm offers a comprehensive, adaptive & scalable approach to identifying the beneficial owners of tokens on-chain. Put simply, Clusters are simply wallets that are connected in any way. If Wallet A has sent funds to Wallet B and they both own Token 'A' this would register as a cluster on the Token A cluster map. This is a 'first-order transaction'. But our clustering algorithm can also identify higher order relationships between wallets, where users have intentionally performed many operations to try and hide their connection, which means we can find relationships between wallets that other tools don't know exist.

We've innovated on this product in three major ways:

1. Live & Scalable Advanced Clustering

Firstly, we've applied our proprietary AI models to create our own clustering scores. This tool allows you to look at any token and determine who the major holders, and how much they really own in real time. As new token transfers and transactions occur on-chain, our clustering algorithm can process this data to refine existing clusters, or create new ones. This allows our users to get the most accurate information in real time, a rarity with clustering tools akin to this one.

2. Higher Order Transaction Handling

Secondly we scan for third and higher-order transactions that relate wallets, which makes it harder to hide the true owners of tokens. We've also developed an infographic that can show you exactly how these clusters are inter-related highlighting the transactions our AI model has picked up on. This results in a more accurate tool able to see past the opaque window of token transactions & wallet relationships on-chain.

Directed graph of transfer interactions with relevant transfers highlighted between token holders & other addresses

3. Data Ingestion

Traditional clustering looks at a narrow view of data on-chain to compile sorted solutions. At the centre of our clustering algorithm, we infuse different data sources employing a combination of network analysis, transaction history & token ownership patterns. This results in our clustering tool being able to capture even the most nuanced of relationships between wallets, which creates a more accurate tool for our users.