AI SECURITY
The Varonis browser extension governs the AI sites your employees use. Block unsanctioned AI, stop sensitive data at the prompt, and quarantine risky sessions.
Trusted by 8,000+ customers
OVERVIEW
The Varonis browser extension lets people keep using the AI tools they rely on. It checks what they type and what comes back, and stops only the sites and data you don't allow.
Find every AI service in use, shadow AI included. Set a default posture, then decide site by site: block what’s risky, audit what needs guardrails, and silence what doesn’t.
Find every AI service in use, shadow AI included. Set a default posture, then decide site by site: block what’s risky, audit what needs guardrails, and silence what doesn’t.
Catch sensitive data in prompts and attachments on AI sites, including ChatGPT, Claude, Gemini, and Copilot. Powered by Varonis data classification.
Catch sensitive data in prompts and attachments on AI sites, including ChatGPT, Claude, Gemini, and Copilot. Powered by Varonis data classification.
The extension blocks sites and prompts on its own. You don't need Cloudflare, Netskope, or another network tool to enforce policy.
The extension blocks sites and prompts on its own. You don't need Cloudflare, Netskope, or another network tool to enforce policy.
Focus on the permissions and exposed data that could increase the impact of a breach.
Bring data discovery, access visibility, and activity monitoring together to support your compliance program.
ACCESS
Set each AI service to Audit, Block, or Silence. Define which tools employees can use, even shadow AI. Blocked sites are stopped in the browser, and every attempt is logged.
PROMPTS
When someone submits a prompt, Atlas scans the text and any attachments for sensitive data. Block the prompt or redact the data before it reaches the AI provider. Blocked prompts stay in the text box, along with a message explaining your policy.
Conversations
Atlas builds a full timeline of each AI session: every prompt, response, upload, and download. Investigate what was asked, what came back, and which guardrails Atlas applied.
QUARANTINE
One risky prompt might be an accident. Two or three is a pattern. When a session crosses your risk threshold, Atlas quarantines it automatically for up to 24 hours or until the session is reviewed.
FAQs
Shadow AI monitoring reads your network and security logs. The browser extension runs in the browser, so it can block a site and inspect what people type.
22 at launch, including ChatGPT, Claude, Gemini, Microsoft Copilot, GitHub Copilot, Perplexity, and Notion AI. New sites are added in days, with no extension update.
Chrome and Edge.
Roll out the extension through the Varonis Data Security Platform. Turn on the browser extension feature on your data plane. Then set your policies in Atlas.
Only when a policy blocks content. Alerts, logs, and quarantines run in the background with no added delay.
No. Prompts and files are evaluated in your own data plane. Captured content is encrypted with your keys.
Yes. Add a custom AI service by URL pattern. Then allow it, audit it, or block it.
