Early access for legal and regulated teams now open Request access →
Login Join Early Access
Browser-based control layer

Control what your team sends to AI.

Ventrin detects risky prompts, removes identifying details locally and lets your team use AI tools safely.

Download for Chrome Join Early Access
Try an example
ChatGPT
Ventrin
How can I help today?
ChatGPT can make mistakes. Check important info.

Works with

  • ChatGPT
  • Claude
  • Gemini
  • Copilot
  • and more
I — The product

See exactly what Ventrin catches.

Three real scenarios, three control modes — all decided in the browser before anything leaves your machine.

chat.openai.com Ventrin
II — How it works

Three steps. Browser-first by design.

  1. 01

    Detect locally

    Ventrin checks prompts in the browser before anything is sent.

  2. 02

    Sanitise intelligently

    Sensitive details are removed or generalised while preserving the user's intent.

  3. 03

    Allow safe use

    Teams keep using AI tools without exposing client data, internal information or confidentiality.

Only sanitised content is ever eligible for AI-assisted rewriting. Confidential information is not submitted to AI tools.

III — Capabilities

Built for secure work.

Instant protection

Works in real time, directly in the browser, with no change to employee workflow.

Built for Chrome

Install as a Chrome extension and start protecting ChatGPT, Claude, and Gemini immediately.

Clear policy control

Set rules by workspace, team, and destination tool. Granular, auditable, calm.

Safe audit trail

Track blocked, warned, and sanitised prompts with secure logging and admin review.

Admin visibility

See what is being caught.

Ventrin logs prompt events, policy triggers, and safe rewrites so admins can understand usage, review false flags, and refine policies over time.

  • Event history
  • Policy hits, by category
  • Team visibility
  • False-flag review queue
Recent events Last 24 hours · 312 events
TimeUserToolActionCategory
14:02:11j.birdChatGPT Blocked Credentials
14:01:48s.millerClaude Sanitised Client identifier
13:58:02r.patelGemini Warned Internal project
13:55:37j.birdChatGPT Sanitised Client identifier
13:51:09t.adekoyaChatGPT Allowed —
Policy: Default · Workspace strictness 2/3
Architecture

Built for confidential work.

  • Prompts are checked locally in real time.
  • Sensitive data is removed or blocked.
  • Only safe input is ever sent onward.
Prompt
Local
Detect
Local
Clean
Send
Browser External AI tool
IV — Who it is for

Legal teams

Stop client matter details, names, and addresses reaching public AI tools — without blocking your team from using them.

Regulated businesses

Meet data-handling obligations on prompt content with browser-level enforcement and full event history.

Security-conscious organisations

Define exactly what types of information may leave the browser, and to which AI destinations, by team and workspace.

V — Plans

Simple plans for data security.

Pick the level of control that fits your team. Upgrade or cancel at any time.

Starter

For individuals and very small teams.

£59 per month, billed monthly
  • Up to 5 users
  • Browser-first prompt protection
  • ChatGPT, Claude, Gemini, Copilot
  • Local-first detection and sanitisation
  • Standard policy templates
  • Email support
Most popular

Team

For teams that need full control.

£199 per month, billed monthly
  • Up to 50 users
  • Everything in Starter, plus:
  • Custom policies by team and tool
  • Document upload protection
  • Full event log and admin dashboard
  • False-flag review workflow
  • Priority support

Enterprise

For organisations with advanced requirements.

Custom Unlimited users · annual contracts
  • Unlimited users
  • Everything in Team, plus:
  • SSO and SCIM provisioning
  • Custom rule packs and vertical policies
  • Optional self-hosted policy verifier
  • Dedicated support and SLA
  • Security and procurement support
Contact Sales

Annual plans save ~17% for teams committing long-term. All plans include the same protection layer — billing only affects price, not access.

Early access pricing. Subject to change as the product evolves.

Common questions.

Does Ventrin block or rewrite prompts?

Both. Lower-risk content is sanitised — identifying details are removed and the user's intent is preserved. Credentials and high-risk content are blocked outright. Admins set the threshold per policy.

Does sensitive information leave the browser?

No. Detection runs locally. Only sanitised content is ever eligible for an optional AI-assisted rewrite, and a server-side verifier confirms nothing identifying remains before that step.

Which AI tools does it support?

ChatGPT, Claude, and Gemini at launch, with more on the roadmap. Anything that runs in the browser is in scope.

Does it work with Chrome?

Yes. Ventrin ships as a Manifest V3 Chrome extension. Edge support follows the same engine.

Can admins set custom policies?

Yes. Define policies by category, severity, and destination tool. Apply at workspace level, with optional team overrides.

What gets logged?

Event metadata (time, user, destination, action, triggered categories) is always logged. Prompt content is logged either fully or masked, depending on the workspace setting.

V — Get started

Let your team use AI safely.

Join Early Access to see how Ventrin fits into your workflow.

Already have an account? Login →

Built for Chrome. Designed for regulated teams.