Most NFT analytics report around fraud. bitsCrunch reports on it: models built specifically to detect wash trading patterns and to identify counterfeit collections copying legitimate art.
Counterfeit detection
Copied collections are a persistent problem — identical artwork minted under a different contract, listed to catch buyers who do not verify the contract address. Detecting them requires image and metadata comparison at scale, which is exactly the kind of task machine learning suits and which no marketplace does thoroughly.
Wash trade detection
Similar to CryptoSlam's approach with different models, flagging suspicious transaction patterns at the collection and wallet level. The internals are not published, which is the standard limitation of proprietary detection systems: you cannot evaluate the false-positive rate.
Coverage and access
Narrower than the general platforms, with a free tier covering core forensic checks and developer APIs used by other products. The company's data feeds appear inside third-party tools more often than users encounter its own interface.
Who should use it
Buyers verifying a collection before purchase, particularly outside the well-known names, and developers needing fraud-detection data feeds.