Skip to content
Sanitized AI Help Center

What is Sanitized AI?

Sanitized AI helps your organization use AI tools such as ChatGPT, Claude, Gemini, and Copilot without leaking sensitive data. It checks each prompt before it’s sent. If the prompt contains something your organization protects, such as a client’s name, a health record, or a password, Sanitized AI stops it and shows you what it found.

You then choose what to do: replace the sensitive parts with safe stand-ins, reword the prompt yourself, or, when your organization allows it, send it anyway. Either way, you keep working.

How it works for you

  1. You type or paste a prompt into an AI tool, as usual.
  2. Before the prompt is sent, Sanitized AI checks it.
  3. If nothing sensitive is found, the prompt is sent and you don’t notice a thing.
  4. If something is found, a warning appears in the corner of the page. It names what it found, explains why it matters, and gives you options.

Files you attach can be checked the same way, if file checking is on for your organization.

What it protects

Your organization decides what counts as sensitive. Common examples:

  • Identity: names, government ID numbers, passport numbers.
  • Contact details: email addresses, phone numbers, home addresses.
  • Financial data: credit card numbers, bank accounts, tax numbers.
  • Health information: health card numbers and other health identifiers.
  • Credentials and secrets: passwords, access keys, and tokens pasted by mistake.
  • Confidential business information: internal identifiers and other business data your organization protects.

Where it works

  • The browser extension protects you on AI websites. It works in Google Chrome, Microsoft Edge, Firefox, and Chromium-based browsers such as Brave.
  • The web dashboard at app.sanitized.ai is where admins manage the organization and where managers see their team’s activity.

Your organization chooses which AI tools are protected. The usual ones include ChatGPT, Claude, Gemini, Microsoft Copilot, and Perplexity.

What your organization can see

Sanitized AI records that something was flagged: which kind of data, on which AI tool, when, and what you did next. It doesn’t store the text of your prompts, and nobody in your organization can read your conversations. For the full picture, see Privacy.

Next steps