ETH Watchtower | Real-Time Blockchain Forensics & Risk Scoring

Every single block, malicious actors deploy scam tokens and factory-generated fraud. Investigators are drowning in noise and flying blind:

• Too Slow: Post-mortem analysis happens after the money is gone.
• Black Boxes: Risk scores give a number, but don't explain why.
• No Visualization: Lists of transactions hide the relationships between deployers.

Sophisticated researchers need more than a file explorer. They need real-time intelligence.

1. Live Monitor

Real-time event streaming. Filter 100,000+ events without lag the millisecond they hit the mempool.

2. Risk Engine

Composite risk score (0-999) based on heuristic weightings. Transparent logistic regression models, not black boxes.

• BurstMint: Risk
• Self-Allocation: Risk
• High-Frequency: Risk

3. Graph Explorer

Visual intelligence. Automatically cluster contracts by bytecode hash and map deployer relationships.

Scammers reuse code. We catch them.

ETH Watchtower automatically clusters contracts by bytecode hash, source code and deployers to visualize the ecosystem.

• Bytecode Pivot: Detect "factory patterns" even if the wallet is fresh.
• Visual Intelligence: Instantly map the web of relationships between deployers.
• Flag Heatmaps: Identify threats that human analysts miss.

Read more about the TUI here.

1. The Hidden Tax Honeypot

Scenario: A token launched with "0% Tax" marketing.

Watchtower Detection: Identified a hidden _taxFee variable set to 99% in the unverified constructor arguments.

Outcome: Flagged as CRITICAL (98/100) 12 seconds after deployment.

2. The "Pepe" Copycat Ring

Scenario: 14 seemingly unrelated tokens launched simultaneously.

Watchtower Detection: Graph Explorer clustered all contracts by bytecode hash and linked them to a single Tornado Cash funded deployer with a history of deploying similar contracts.

Outcome: Identified as a coordinated rugpull ring before liquidity was added.

React + TypeScript + SQLite
A responsive frontend backed by a robust streaming backend service.

Memory-Efficient Log Ingester
Handles massive datasets.

High-Throughput Go API
Custom backend services written in Go for historical data aggregation and cross-chain signal normalization.

Epoch-Aware Analysis
Parses historical boundaries (The Merge, Shanghai, London) to apply context-specific heuristics rather than generic scanning.

Local-First Forensics
Case management and investigation notes are stored locally. We don't see your alpha.

Modular Design
Decoupled 'Analytics' allows plugging in new risk models instantly.

Blockchains are transparent machines. Security intelligence should be transparent too.

Rather than hiding analysis behind proprietary walls, Watchtower treats the chain as a public library of code, incentives, and human creativity — sometimes brilliant, sometimes malicious, always fascinating.

By illuminating how contracts actually behave, the tool contributes to a safer and more self-aware cryptoeconomic ecosystem.