AI security, without the stress

Let your team use AI. Keep their secrets off it.

Inveil runs quietly on each machine and catches personal info before it leaves the device. It blocks the request or strips the data out, and your team keeps working.

Book a call See how it works

Works with the AI tools your team already uses

  • ChatGPT
  • Claude
  • Claude Code
  • Gemini
  • Copilot
  • Cursor

Checked on the device. Clean by the time it leaves.

Personal information never leaves your machine. We catch it before it reaches the AI.

Your employee writes a prompt

They use ChatGPT, Claude, Cursor or any tool you have chosen, exactly as they do today.

Inveil checks it on the device

A small local proxy looks for emails, SSNs, card numbers and API keys. Nothing is sent anywhere to be checked.

Only clean requests reach the AI

Risky ones are blocked, or redacted and sent on. Either way, the employee is told what happened.

Two modes. No new habits.

Choose block or redact for each rule. Employees see a plain note in the tool they are already using.

Block: the request stops on the machine and nothing is sent.
Redact: the email is removed, the request completes, and the reply still arrives.

Your rules, your call.

Start with the built-in detectors, then add what matters to your company. Every rule can block or redact.

  • Built-in detectors

    SSNs, card numbers, email addresses and API keys are covered from day one, in text, images and attachments. Turn each one on or off.

  • Rules of your own

    Add specific email addresses or whole domains, words and phrases like a project codename, or patterns such as your customer IDs.

  • Block or redact, rule by rule

    Stop a request outright for the sensitive things. Strip the data and let the rest go through. Change it whenever you like.

Files and screenshots too.

People paste screenshots and attach documents as often as they type. Inveil checks those as well, before anything is sent.

  • Read on the device

    Images, screenshots and PDFs are read with built-in OCR on the machine. No AI model and no outside API is involved.

  • Filenames are checked too

    A name like jane-doe-ssn-tax-return.pdf can give away as much as the file itself, so names get the same check.

  • Same rules, block or redact

    What you set for prompts applies to files. Stop the request, or strip the details and let it go through.

Everything you need. Nothing to babysit.

Stays on the machine

The check runs locally on each device, so personal information is caught before it goes anywhere.

You pick the providers

Choose which AI tools Inveil watches. Add or change them whenever your team does.

Finds the usual suspects

Detects emails, SSNs, card numbers and API keys in prompts, images and attachments. Images are read with on-device OCR, not an AI model.

Block or redact, per rule

Set each rule, built-in or your own, to stop the request or strip the data and let it complete. Employees are told exactly what was removed.

No change to how people work

No new app to learn and no extra steps. People keep using the tools they like.

Rolls out to the whole team

Set it up once for everyone. Every event is logged against the person it belongs to.

See your AI usage, and what it costs.

One calm view of how your company uses AI, with a record of every block and redaction.

  • Usage by employee, team and provider
  • Token counts and estimated cost, month to date
  • An activity feed of every block and redaction, per person
  • Know who triggered a rule, and what each request cost

Questions, answered

Make AI at work simple to trust.

Book a short call. We will show you Inveil on a real prompt and answer anything you like.

Book a call