Model Accuracy
—
Measured on the test set
Dataset Size
—
Labeled URLs · 1:1 balanced
Avg Latency
—ms
Per-URL inference time
URLs Scanned
0
This session
Live Feed
Initializing threat feed…
Analyze a URL
Quick test samples:
Extracting 22 signals, classifying…
Enter a URL above to begin threat analysis
Phishing Risk Score
0%
SAFE
Safe
Analyzed URL
—
Confidence
—
Latency
—
Classification
—
Feature Breakdown 22 signals
Scan a URL to view extracted signals
Feature Comparison — Legitimate vs Phishing
Signal Categories
Length Metrics
url_length · hostname_length · path_length · query_length
Character Analysis
dots · hyphens · underscores · slashes · digit_ratio · letter_ratio
Structural
has_ip_address · subdomain_count · tld_length · has_https
Risk Indicators
suspicious TLD · shortener · @ symbol · double-slash · hex encoding · phishing keywords
Shannon Entropy
Detects obfuscated, randomized character patterns common in phishing URLs
Classifier Benchmark Results 5-fold CV · 2,000 URL corpus
Deployed Model
Selected
—
Accuracy
—
CV F1 (5-fold)
—
Precision
—
Recall
—
Confusion Matrix
Pred: Legit
Pred: Phish
Act: Legit
—
—
Act: Phish
—
—
Detection Rate
—
False Alarm
—
Batch URL Input
One URL per line. Classifies up to 50 URLs in a single request.
Results
·
No results yet — enter URLs and click Run.
Training Dataset Overview
Legitimate
—
Phishing
—
Total
—
Features
22
Dataset Insights
Avg Length — Legitimate
—
Avg Length — Phishing
—
Suspicious TLD Usage
Legit: —
Phish: —
IP-Based URLs
Legit: —
Phish: —
Data Sources
PhishTank · OpenPhish · Kaggle URL Datasets (locally curated)
Training Data Sample
Loading preview…
Scan History
No scans yet. Use the URL Scanner tab.