Identra Logo
IDENTRA
IDENTRA GUARDA browser extension

Own the browser.Own the AI in it.

Identra Guard finds the AI apps and accounts people use, checks prompts on the device before they're sent, scores every extension, catches agents driving the browser and guards logins against phishing.

See account-aware access
chatgpt.com
maya · work

Draft a renewal email for this customer. Card on file: card number. API key: secret key.

AI PROMPT
PROMPT CHECKED ON THIS DEVICE

§ 01Account-aware access

Work AI accounts in. Personal ones redirected.

Allow, redirect to your company AI workspace, or block, decided by the account someone is signed in with. For ChatGPT, Gemini, Claude, Perplexity and Grok.

  • Discovers AI apps and the accounts people use with them.

  • Steers personal accounts to the approved workspace instead of a dead end.

  • Decisions tied to the person's identity.

AI APP
SIGNED IN WITH
POLICY
ALLOWCHATGPT · WORK ACCOUNT

Allowed

§ 02Prompt protection

AI prompts checked on the device. Before they're sent.

An on-device AI classifier checks each prompt before it is sent, then allowed, alerted, masked or blocked by policy. Prompt content stays on the device by default.

BY POLICY
  • ALLOW
  • ALERT
  • MASKreplace and send
  • BLOCK

Sensitive data

Cards, tokens, SSNs, API keys, credentials and more.

Content and intent

Sensitive content and risky intent, like a request to bypass a security control.

Uploads

Uploads to AI checked for sensitive files. File uploads to AI can be blocked by policy.

Clipboard

Clipboard controls and secret detection in pastes.

§ 03Extensions

Every extension, scored.

Every extension gets a trust score. Extensions that read sign-in sessions are flagged, and risky ones can be disabled by policy.

  • WHY IT WAS FLAGGED
    reads sign-in sessions

§ 04Agents in the browser

When an agent drives the browser, you'll know.

Identra Guard tells you when an AI agent, not a person, is driving the browser, and can block it by policy.

WHAT GUARD SAW
  • Agent driving the pageWatching
  • File attached by an agentWatching
  • Agent sending the messageWatching
POLICY
READY

§ 05Credential phishing

AI runs on your sessions. Guard the login.

Guard spots fake login pages and the lures that steal sessions and credentials.

Lookalike logins

Lookalike login pages, typosquats and phishing kits caught.

QR and device‑code lures

QR and device-code phishing caught at the page.

Risky sites

Corporate credentials typed on risky sites are flagged.

Password reuse

Reused and breached passwords flagged.

OAuth consent

Consent screens for Microsoft and Google mediated.

Fake CAPTCHA

Fake-CAPTCHA command lures detected and alerted.

§ 06Resource policy

AI apps and SaaS, policy per domain group.

Block, redirect, banner, watermark, or control copy and paste, for each group of domains.

DOMAIN GROUP
  • Block
  • Redirect
  • Banner
  • Watermark
  • Copy and paste

Browser activity lands on the person's timeline.

See one identity, one timeline
QUESTIONS

Questions about Guard.

Which browsers does Identra Guard support?

Identra Guard is a browser extension for Chrome, Edge, Firefox and Safari.

Does Guard send prompts to the cloud for inspection?

No. An on-device AI classifier checks prompts before send, and prompt content stays on the device by default.

Can Guard tell a work account from a personal one?

Yes. It discovers the account used with each AI app and applies allow, redirect or block by signed-in account for ChatGPT, Gemini, Claude, Perplexity and Grok.

What happens to risky extensions?

Guard gives each extension a trust score, flags those that read sign-in sessions, and can disable risky ones by policy.

What does Guard collect?

See the data collection notice and the Identra Guard privacy policy.

KEEP READING

COMPARE

LEARN THE TERMS

The browser AI terms, defined.

Browse the AI security glossary
BOOK A WALKTHROUGH

Put Guard in front of your AI.

A walkthrough with a security engineer.