Specter

Most websites silently load dozens of third-party scripts the moment you visit them. Ad networks. Behavioral trackers. Session replay tools recording your every click and scroll. Browser fingerprinters collecting your device attributes without asking.
Specter makes all of it visible.
What Specter does:
Specter intercepts every network request your browser makes and classifies it in real time using a locally-running XGBoost ML model — trained on 25,290 real browsing requests with 96.2% accuracy and a weighted F1 of 0.967. No data ever leaves your machine. No server. No proxy. No account required.
Categories detected:
— Behavioral trackers
— Fingerprinting scripts
— Ad networks
— Analytics
— Legitimate third-party
Features:
— Live request feed with category badges, confidence scores, and per-request explainability
— Privacy score (0–100) per site, updated in real time
— D3.js timeline showing request bursts relative to page load
— Session history with per-site breakdowns, sortable by date, tracker count, or privacy score
— Full JSON export or plain-text report copy for any session
— Optional VirusTotal integration for domain reputation lookups (opt-in, domain names only)
How it works:
The ML classifier extracts 27 features per request — URL structure, response size, header patterns, domain reputation, path semantics — and classifies each request in under 50ms. Everything runs locally via ONNX Runtime. A rule-based fallback handles session replay detection and cold-start scenarios.
Privacy by design:
Specter does not collect, transmit, or store any data externally. All session data lives in your browser's IndexedDB and is deleted on your schedule.
Specter is an auditor, not a blocker. It doesn't interfere with your browsing — it just shows you the truth about it.