AI-Powered Smart Contract Intelligence

KNOW BEFORE YOU BUY IT.

Paste any Solidity contract or Ethereum token address. RugGuard runs 53 risk signals through our XGBoost model and returns an audit-grade verdict in under 3 seconds.

// supports ETH token addresses and raw Solidity — no account needed

RUNNING 53-FEATURE ANALYSIS...

96.1%
Model Accuracy
99.6%
Precision Score
0.985
AUC-ROC Score
53
Risk Signals
Detection Coverage

7 CATEGORIES.
53 SIGNALS.

Every detection maps to a confirmed real-world rug pull pattern from 2020–2025.

🛡️
Fund-Drain Detection
Identifies functions that withdraw the full contract balance, drain ERC-20 tokens, or call selfdestruct — including hidden paths in complex call chains.
Critical
👑
Ownership Abuse
Detects hidden owner addresses, hash-based access control, proxy traps, and contracts that cannot renounce ownership — even after claiming to.
Critical
🍯
Honeypot Mechanics
Flags contracts where buying is enabled but selling is blocked — via explicit revert(), gas restrictions, or owner-only transfer permissions.
Critical
⏱️
Time-Based Rug Patterns
Catches delayed drain mechanics using block.timestamp — contracts appearing safe at launch but unlocking malicious functions after a set period.
High
🔄
Proxy & Upgradeability Risk
Detects upgradeable proxies without timelocks — where the owner can silently swap the entire contract logic after launch.
High
🔍
Code Obfuscation
Flags assembly-level backdoors, keccak256 address hiding, encoded parameter obfuscation, and suspiciously low documentation ratios.
High
💧
Liquidity Risk Signals
Checks for LP lock status, unrestricted liquidity removal, burn mechanism verification, and vesting schedule presence before you invest.
High
⚡
tx.origin & Fallback Traps
tx.origin phishing in economic contexts, dangerous logic inside fallback() handlers, and gas-based execution restrictions that block selling.
Critical
Model Performance

NUMBERS YOU CAN
ACTUALLY DEFEND.

XGBoost trained on 2,474 labeled contracts. 5-fold cross-validation. CV std 0.0043 — not a fluke.

96.1%
Accuracy
99.6%
Precision
96.4%
Recall
0.980
F1 Score
5-fold cross-validation
cv_accuracy: 0.9716 ± 0.0043
cv_f1: 0.9852 ± std
cv_auc: 0.9897

Low std deviation = consistent, not overfitting.
// confusion matrix — holdout test set
474
True Positive
Rugs caught
2
False Positive
Safe flagged
17
False Negative
Rug missed
0
True Negative
Safe cleared
Product Roadmap

BUILT TO GROW
WITH THE THREAT.

Phase 01
🌐
Web Analyzer
Paste contract code or a token address. XGBoost + 53-feature semantic analysis in under 3 seconds. SHAP explainability shows exactly why.
Live
Phase 02
🧩
Chrome Extension
Risk overlays injected directly on Etherscan, Uniswap, and DexScreener. See the score before you click buy — zero friction.
In Development
Phase 03
📱
Mobile App
Wallet monitoring, push alerts for newly deployed high-risk contracts, and on-the-go analysis for any chain.
Planned
Phase 04
⚡
REST API
Integrate RugGuard risk scores directly into your dApp, DEX interface, or trading bot. Per-address caching, webhook support.
In Development
Phase 05
🔔
Wallet Alerts
Connect your wallet and get notified the moment a token you hold shows critical risk signals — before the dump happens.
Planned
Phase 06
📊
Risk Dashboard
Portfolio-level risk overview. Historical analysis. Shareable audit reports. Team verification scores. The full picture.
Planned

START ANALYZING.
IT'S FREE.

No account. No credit card. Just paste a contract and know.

Analyze a contract View on GitHub →