Attack prevention, data loss protection, threat detection and compliance — all running locally in your browser. No agents, no proxies, no data leaving your machine.
🧠
AI Manipulation & Prompt Attack Protection Protect+ New
Scores the intent of every prompt before it reaches any AI model. Catches direct injection, jailbreak personas, indirect RAG attacks, encoded payloads, privilege escalation and multi-step manipulation — in 7 languages.
🔍
Real-time prompt scanning Free
As-you-type detection across all 45+ supported platforms. Risk score 0–100. Scans for PII, secrets, financial data, company confidentials and attack patterns simultaneously.
📋
Clipboard protection Protect
Intercepts paste and copy events in real time. Detects sensitive data before it enters any AI tool. The only extension scanning both the prompt input and the clipboard event stream.
📎
File upload inspection Protect
Scans 20+ file types on input change and drag-and-drop — including txt, csv, json, js, py, sql and best-effort PDF/DOCX extraction. Warns before the file reaches the AI model.
🖨
Print interception Protect
Fires on Ctrl+P and window.print(). Scans page content for sensitive data before it exits the browser — a last-mile DLP capability no API proxy can provide.
⚠️
Malware inspection New Protect
Detects malware filenames, encoded payloads, reverse shell patterns, dropper language, registry persistence hints and binary obfuscation in prompts and file uploads.
🎣
Phishing & URL protection New Protect
Flags credential harvesting language, brand impersonation typosquats, URL shorteners, homoglyph domains and phishing kit terminology before they reach an AI tool.
🤖
Fake CAPTCHA detection New Protect
Stops ClickFix attacks that trick users into pasting malicious commands. Detects "press Win+R", clipboard payload delivery and social-engineering CAPTCHA bypass language.
🎯
Adaptive risk policy New Protect+
Elevates risk scores based on contextual signals: urgency cues, secrecy requests, authority impersonation, evasion attempts and multi-step exfiltration chains.
🔌
BYOMCP Integrations New Enforce
Forward detection events to any MCP server, SIEM or webhook. Supports Splunk, Datadog, Microsoft Sentinel, Slack, MS Teams. Formats: JSON, CEF, LEEF, OCSF 1.1, custom.
🧠
AI Model Blocker New Enforce
Blocks Chrome Gemini Nano, WebNN local ML inference and Transformers.js/ONNX model loads. Prevents sites from silently downloading gigabytes of model weights to your device.
📊
Risk analytics & heatmap Protect+
7×24 detection heatmap, 7-day trend analytics, detection source breakdown and 250-event searchable log — all stored locally in your browser.
📤
SIEM & compliance exports Enforce
Export structured JSON for Splunk, Datadog or Microsoft Sentinel. Pre-formatted HIPAA, GDPR and PCI-DSS compliance snapshots generated locally for audit evidence.
