Fraud Detector¶

Key Takeaways
- Predictive fraud scoring: identifies wallets likely to commit fraud from behavioural patterns, not just wallets already caught in databases
- 98% accuracy on Ethereum; sub-5% false positive rate; covers 8 blockchains; free to use with no signup required
- AML and sanctions screening built in: OFAC SDN, EU Consolidated Sanctions List, UN Security Council, plus mixer exposure and darknet connections
- Results in under 100ms for cached addresses; also available via Telegram bot

Most fraud detection tools tell you about wallets that have already been caught. ChainAware's Fraud Detector tells you about wallets that haven't been caught yet.

By analysing on-chain behavioural patterns across 20M+ wallet profiles, it predicts fraud risk with 98% accuracy - before any crime occurs, before any report is filed, before anyone else has flagged the address. The result is a fraud probability score from 0 to 1, delivered in under 100ms.

Free to use. No signup. No wallet connection needed.

Check a Wallet's Fraud Score →

Predictive vs. Forensic: Why It Matters¶

Most crypto security tools are forensic - they maintain databases of addresses that have already been confirmed as fraudulent. If an address isn't in the database, it gets a clean bill of health, even if it's about to scam you tomorrow.

ChainAware is predictive. The AI analyses the behavioural history of a wallet - its transaction patterns, counterparty relationships, protocol usage, timing behaviours - and identifies the signatures of fraud before it manifests. This catches:

New scammer wallets that haven't been reported yet
Sophisticated actors who rotate through clean-looking addresses
Wallets with emerging risky patterns that haven't crossed a threshold yet

The difference in practice: A forensic tool says "this address hasn't been caught before." ChainAware says "this address behaves like 94% of wallets that went on to commit fraud."

What the Fraud Score Means¶

The score runs from 0 to 1:

Score Range	Interpretation
0.00 - 0.25	Low risk - behavioural patterns consistent with legitimate activity
0.26 - 0.50	Moderate - some anomalous signals; proceed with caution
0.51 - 0.75	High risk - significant fraud indicators present
0.76 - 1.00	Very high risk - strong fraud pattern match; avoid interaction

Addresses already confirmed in forensic databases as scam addresses are automatically scored at 1.0 regardless of the AI model's output.

How It Works¶

ChainAware's fraud model evaluates wallets across ten behavioural parameters - the same ones that underpin the full Wallet Auditor:

Transaction patterns - frequency, value distribution, timing, and counterparty diversity
Protocol behaviour - which protocols are used, in what order, and with what cadence
Network connections - relationships to known bad actors, mixers, sanctioned addresses
Experience signals - is this wallet behaving consistently with its stated history?
Anomaly indicators - sudden behavioural shifts, wash trading patterns, rapid fund movement

The model is an ensemble - multiple specialised algorithms vote on fraud likelihood, with consensus reducing false positives. It retrains daily on new fraud examples, adapting to evolving attack techniques faster than rule-based systems can be updated.

False positive rate: ChainAware targets sub-5% false positives, far below the 30-70% rates typical of traditional rule-based systems.

AML and Sanctions Monitoring - Fully Integrated¶

AML and sanctions screening are built directly into the Fraud Detector - not a separate check. Every fraud score includes:

Sanctions screening - OFAC SDN, EU Consolidated Sanctions List, UN Security Council lists
Mixer and tumbler exposure - Tornado Cash and known mixing services
Darknet market connections - direct and indirect links to illicit fund clusters
Hack and exploit proceeds - funds traceable to known protocol hacks or exit scams
Counterparty network analysis - flagged relationships traced through the transaction graph

The full Wallet Auditor exposes these as a dedicated AML and Sanctions Monitoring parameter with structured flag details for compliance use.

Fraud Detection vs. AML: What's the Difference?¶

These are complementary signals - both are included in every check:

Fraud Detection - predicts whether a wallet is likely to commit fraud in the future, based on behavioural patterns. Answers: is this wallet dangerous to interact with?

AML (Anti-Money Laundering) - verifies whether funds passing through a wallet have legal origins, checking against sanctions lists and known crime proceeds. Answers: are these funds clean?

A wallet can have clean AML status but high fraud risk (a new scammer with no prior history). Or it can have AML flags but low forward-looking fraud risk (a historical connection that's no longer active). Both signals are returned together - you don't have to choose.

How to Use the Fraud Detector¶

Go to chainaware.ai/fraud-detector
Select the blockchain (Ethereum, BNB Smart Chain, Polygon, Base, Solana, TON, Tron, or Haqq)
Paste any wallet address - ENS and Unstoppable Domains names are supported
Click Calculate

Results are returned instantly from cache for previously analysed addresses. For addresses not yet in the system, recalculation takes 3-4 seconds for wallets with large transaction histories.

Note: The detector works on regular wallet addresses only. For contract addresses (tokens, pools, protocols), use the Rug Pull Detector instead.

When the Detector Can't Score a Wallet¶

Two situations produce no score:

Contract addresses - the AI is trained on wallet behaviour, not contract logic. Paste a contract and ChainAware will redirect you to the Rug Pull Detector automatically.
New wallets with fewer than 10-15 transactions - there isn't enough behavioural data to establish a reliable pattern. A brand new wallet with no history scores as unknown, not safe.

A wallet with no score is not necessarily trustworthy - it may simply be too new to have developed a detectable pattern.

What You Can Use It For¶

Before sending funds - Check any address before transferring cryptocurrency. Even a 10-second fraud check can prevent a $2,600 average loss.

Someone's offering you an investment deal - Anyone pitching a crypto opportunity has a wallet. Run it before you commit.

A stranger messaged you on Telegram or Discord - Crypto scams disproportionately start on messaging platforms. A quick fraud check tells you whether the wallet they reference is legitimate.

Verifying a payment address - Address substitution attacks replace legitimate receiving addresses. Audit the final destination before signing any transaction.

Screening NFT sellers or buyers - Before a high-value NFT trade, check the counterparty's fraud score.

DeFi liquidity partners - Before adding joint liquidity with another party, verify their on-chain reputation.

Access via Bot¶

Run fraud checks without leaving your community:

Telegram: /fraudcheck [address] - Add ChainAware Bot

Supported Networks¶

Ethereum, BNB Smart Chain, Polygon, Base, Solana, TON, Tron, and Haqq.

API Access¶

Direct API access is available at swagger.chainaware.ai with an API key from your profile dashboard. An active subscription is required for API calls - see chainaware.ai/pricing.

Frequently Asked Questions¶

How is a predictive fraud score different from a blocklist?
A blocklist flags wallets already confirmed as fraudulent - it only catches wallets that have been caught before. ChainAware analyses the behavioural patterns of a wallet and identifies signatures associated with fraud before it is reported. A new scammer with no prior history gets a high fraud score based on how their wallet behaves, not whether they appear in a database.

What is the false positive rate?
ChainAware targets under 5% false positives - far below the 30-70% false positive rates typical of rule-based forensic systems. High false positive rates flag legitimate users, which creates friction and damages conversion.

Can it score wallets with no history?
Not reliably. Wallets with fewer than 10-15 transactions do not have enough behavioural data. A brand-new wallet scores as unknown, not safe - treat it accordingly for large transfers.

Does it include sanctions screening?
Yes. AML and sanctions screening are built into every fraud check. Every result includes OFAC SDN, EU Consolidated Sanctions List, and UN Security Council status, plus mixer exposure, darknet connections, and hack proceeds links.

Is it free?
Yes - free at chainaware.ai with no signup required. Enterprise API for high-volume screening is available at swagger.chainaware.ai.